diff options
108 files changed, 18785 insertions, 0 deletions
diff --git a/build/FindAnacondaEnvironment.cmake b/build/FindAnacondaEnvironment.cmake new file mode 100644 index 0000000..6475128 --- /dev/null +++ b/build/FindAnacondaEnvironment.cmake @@ -0,0 +1,154 @@ +# Copyright 2017 Edoardo Pasca +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +# #.rst: +# FindAnacondaEnvironment +# -------------- +# +# Find Python executable and library for a specific Anaconda environment +# +# This module finds the Python interpreter for a specific Anaconda enviroment, +# if installed and determines where the include files and libraries are. +# This code sets the following variables: +# +# :: +# PYTHONINTERP_FOUND - if the Python interpret has been found +# PYTHON_EXECUTABLE - the Python interpret found +# PYTHON_LIBRARY - path to the python library +# PYTHON_INCLUDE_PATH - path to where Python.h is found (deprecated) +# PYTHON_INCLUDE_DIRS - path to where Python.h is found +# PYTHONLIBS_VERSION_STRING - version of the Python libs found (since CMake 2.8.8) +# PYTHON_VERSION_MAJOR - major Python version +# PYTHON_VERSION_MINOR - minor Python version +# PYTHON_VERSION_PATCH - patch Python version + + + +function (findPythonForAnacondaEnvironment env) + if (WIN32) + file(TO_CMAKE_PATH ${env}/python.exe PYTHON_EXECUTABLE) + elseif (UNIX) + file(TO_CMAKE_PATH ${env}/bin/python PYTHON_EXECUTABLE) + endif() + + + message("findPythonForAnacondaEnvironment Found Python Executable" ${PYTHON_EXECUTABLE}) + ####### FROM FindPythonInterpr ######## + # determine python version string + if(PYTHON_EXECUTABLE) + execute_process(COMMAND "${PYTHON_EXECUTABLE}" -c + "import sys; sys.stdout.write(';'.join([str(x) for x in sys.version_info[:3]]))" + OUTPUT_VARIABLE _VERSION + RESULT_VARIABLE _PYTHON_VERSION_RESULT + ERROR_QUIET) + if(NOT _PYTHON_VERSION_RESULT) + string(REPLACE ";" "." _PYTHON_VERSION_STRING "${_VERSION}") + list(GET _VERSION 0 _PYTHON_VERSION_MAJOR) + list(GET _VERSION 1 _PYTHON_VERSION_MINOR) + list(GET _VERSION 2 _PYTHON_VERSION_PATCH) + if(PYTHON_VERSION_PATCH EQUAL 0) + # it's called "Python 2.7", not "2.7.0" + string(REGEX REPLACE "\\.0$" "" _PYTHON_VERSION_STRING "${PYTHON_VERSION_STRING}") + endif() + else() + # sys.version predates sys.version_info, so use that + execute_process(COMMAND "${PYTHON_EXECUTABLE}" -c "import sys; sys.stdout.write(sys.version)" + OUTPUT_VARIABLE _VERSION + RESULT_VARIABLE _PYTHON_VERSION_RESULT + ERROR_QUIET) + if(NOT _PYTHON_VERSION_RESULT) + string(REGEX REPLACE " .*" "" _PYTHON_VERSION_STRING "${_VERSION}") + string(REGEX REPLACE "^([0-9]+)\\.[0-9]+.*" "\\1" _PYTHON_VERSION_MAJOR "${PYTHON_VERSION_STRING}") + string(REGEX REPLACE "^[0-9]+\\.([0-9])+.*" "\\1" _PYTHON_VERSION_MINOR "${PYTHON_VERSION_STRING}") + if(PYTHON_VERSION_STRING MATCHES "^[0-9]+\\.[0-9]+\\.([0-9]+)") + set(PYTHON_VERSION_PATCH "${CMAKE_MATCH_1}") + else() + set(PYTHON_VERSION_PATCH "0") + endif() + else() + # sys.version was first documented for Python 1.5, so assume + # this is older. + set(PYTHON_VERSION_STRING "1.4" PARENT_SCOPE) + set(PYTHON_VERSION_MAJOR "1" PARENT_SCOPE) + set(PYTHON_VERSION_MINOR "4" PARENT_SCOPE) + set(PYTHON_VERSION_PATCH "0" PARENT_SCOPE) + endif() + endif() + unset(_PYTHON_VERSION_RESULT) + unset(_VERSION) + endif() + ############################################### + + set (PYTHON_EXECUTABLE ${PYTHON_EXECUTABLE} PARENT_SCOPE) + set (PYTHONINTERP_FOUND "ON" PARENT_SCOPE) + set (PYTHON_VERSION_STRING ${_PYTHON_VERSION_STRING} PARENT_SCOPE) + set (PYTHON_VERSION_MAJOR ${_PYTHON_VERSION_MAJOR} PARENT_SCOPE) + set (PYTHON_VERSION_MINOR ${_PYTHON_VERSION_MINOR} PARENT_SCOPE) + set (PYTHON_VERSION_PATCH ${_PYTHON_VERSION_PATCH} PARENT_SCOPE) + message("My version found " ${PYTHON_VERSION_STRING}) + ## find conda executable + if (WIN32) + set (CONDA_EXECUTABLE ${env}/Script/conda PARENT_SCOPE) + elseif(UNIX) + set (CONDA_EXECUTABLE ${env}/bin/conda PARENT_SCOPE) + endif() +endfunction() + + + +set(Python_ADDITIONAL_VERSIONS 3.5) + +find_package(PythonInterp) +if (PYTHONINTERP_FOUND) + + message("Found interpret " ${PYTHON_EXECUTABLE}) + message("Python Library " ${PYTHON_LIBRARY}) + message("Python Include Dir " ${PYTHON_INCLUDE_DIR}) + message("Python Include Path " ${PYTHON_INCLUDE_PATH}) + + foreach(pv ${PYTHON_VERSION_STRING}) + message("Found interpret " ${pv}) + endforeach() +endif() + + + +find_package(PythonLibs) +if (PYTHONLIB_FOUND) + message("Found PythonLibs PYTHON_LIBRARIES " ${PYTHON_LIBRARIES}) + message("Found PythonLibs PYTHON_INCLUDE_PATH " ${PYTHON_INCLUDE_PATH}) + message("Found PythonLibs PYTHON_INCLUDE_DIRS " ${PYTHON_INCLUDE_DIRS}) + message("Found PythonLibs PYTHONLIBS_VERSION_STRING " ${PYTHONLIBS_VERSION_STRING} ) +else() + message("No PythonLibs Found") +endif() + + + + +function(findPythonPackagesPath) + execute_process(COMMAND ${PYTHON_EXECUTABLE} -c "from distutils.sysconfig import *; print (get_python_lib())" + RESULT_VARIABLE PYTHON_CVPY_PROCESS + OUTPUT_VARIABLE PYTHON_STD_PACKAGES_PATH + OUTPUT_STRIP_TRAILING_WHITESPACE) + #message("STD_PACKAGES " ${PYTHON_STD_PACKAGES_PATH}) + if("${PYTHON_STD_PACKAGES_PATH}" MATCHES "site-packages") + set(_PYTHON_PACKAGES_PATH "python${PYTHON_VERSION_MAJOR_MINOR}/site-packages") + endif() + + SET(PYTHON_PACKAGES_PATH "${PYTHON_STD_PACKAGES_PATH}" PARENT_SCOPE) + +endfunction() + + diff --git a/build/run.sh b/build/run.sh new file mode 100644 index 0000000..a8e5555 --- /dev/null +++ b/build/run.sh @@ -0,0 +1,19 @@ +#!/bin/bash +echo "Building CCPi-regularisation Toolkit using CMake" +# rm -r build +# Requires Cython, install it first: +# pip install cython +# mkdir build +cd build/ +make clean +# install Python modules only without CUDA +cmake ../ -DBUILD_PYTHON_WRAPPER=ON -DBUILD_MATLAB_WRAPPER=OFF -DBUILD_CUDA=OFF -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=./install +# install Python modules only with CUDA +# cmake ../ -DBUILD_PYTHON_WRAPPER=ON -DBUILD_MATLAB_WRAPPER=OFF -DBUILD_CUDA=ON -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=./install +make install +# cp install/lib/libcilreg.so install/python/ccpi/filters +cd install/python +export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:../lib +# spyder +# one can also run Matlab in Linux as: +# PATH="/path/to/mex/:$PATH" LD_LIBRARY_PATH="/path/to/library:$LD_LIBRARY_PATH" matlab diff --git a/docs/data/SinoInpaint.mat b/docs/data/SinoInpaint.mat Binary files differnew file mode 100644 index 0000000..d748fb4 --- /dev/null +++ b/docs/data/SinoInpaint.mat diff --git a/Wrappers/Python/conda-recipe/lena_gray_512.tif b/docs/data/lena_gray_512.tif Binary files differindex f80cafc..f80cafc 100644 --- a/Wrappers/Python/conda-recipe/lena_gray_512.tif +++ b/docs/data/lena_gray_512.tif diff --git a/docs/demos/demoMatlab_3Ddenoise.m b/docs/demos/demoMatlab_3Ddenoise.m new file mode 100644 index 0000000..0c331a4 --- /dev/null +++ b/docs/demos/demoMatlab_3Ddenoise.m @@ -0,0 +1,178 @@ +% Volume (3D) denoising demo using CCPi-RGL +clear; close all +Path1 = sprintf(['..' filesep 'mex_compile' filesep 'installed'], 1i); +Path2 = sprintf(['..' filesep '..' filesep '..' filesep 'data' filesep], 1i); +Path3 = sprintf(['..' filesep 'supp'], 1i); +addpath(Path1); +addpath(Path2); +addpath(Path3); + +N = 512; +slices = 7; +vol3D = zeros(N,N,slices, 'single'); +Ideal3D = zeros(N,N,slices, 'single'); +Im = double(imread('lena_gray_512.tif'))/255; % loading image +for i = 1:slices +vol3D(:,:,i) = Im + .05*randn(size(Im)); +Ideal3D(:,:,i) = Im; +end +vol3D(vol3D < 0) = 0; +figure; imshow(vol3D(:,:,15), [0 1]); title('Noisy image'); + + +lambda_reg = 0.03; % regularsation parameter for all methods +%% +fprintf('Denoise a volume using the ROF-TV model (CPU) \n'); +tau_rof = 0.0025; % time-marching constant +iter_rof = 300; % number of ROF iterations +tic; u_rof = ROF_TV(single(vol3D), lambda_reg, iter_rof, tau_rof); toc; +energyfunc_val_rof = TV_energy(single(u_rof),single(vol3D),lambda_reg, 1); % get energy function value +rmse_rof = (RMSE(Ideal3D(:),u_rof(:))); +fprintf('%s %f \n', 'RMSE error for ROF is:', rmse_rof); +figure; imshow(u_rof(:,:,7), [0 1]); title('ROF-TV denoised volume (CPU)'); +%% +% fprintf('Denoise a volume using the ROF-TV model (GPU) \n'); +% tau_rof = 0.0025; % time-marching constant +% iter_rof = 300; % number of ROF iterations +% tic; u_rofG = ROF_TV_GPU(single(vol3D), lambda_reg, iter_rof, tau_rof); toc; +% rmse_rofG = (RMSE(Ideal3D(:),u_rofG(:))); +% fprintf('%s %f \n', 'RMSE error for ROF is:', rmse_rofG); +% figure; imshow(u_rofG(:,:,7), [0 1]); title('ROF-TV denoised volume (GPU)'); +%% +fprintf('Denoise a volume using the FGP-TV model (CPU) \n'); +iter_fgp = 300; % number of FGP iterations +epsil_tol = 1.0e-05; % tolerance +tic; u_fgp = FGP_TV(single(vol3D), lambda_reg, iter_fgp, epsil_tol); toc; +energyfunc_val_fgp = TV_energy(single(u_fgp),single(vol3D),lambda_reg, 1); % get energy function value +rmse_fgp = (RMSE(Ideal3D(:),u_fgp(:))); +fprintf('%s %f \n', 'RMSE error for FGP-TV is:', rmse_fgp); +figure; imshow(u_fgp(:,:,7), [0 1]); title('FGP-TV denoised volume (CPU)'); +%% +% fprintf('Denoise a volume using the FGP-TV model (GPU) \n'); +% iter_fgp = 300; % number of FGP iterations +% epsil_tol = 1.0e-05; % tolerance +% tic; u_fgpG = FGP_TV_GPU(single(vol3D), lambda_reg, iter_fgp, epsil_tol); toc; +% rmse_fgpG = (RMSE(Ideal3D(:),u_fgpG(:))); +% fprintf('%s %f \n', 'RMSE error for FGP-TV is:', rmse_fgpG); +% figure; imshow(u_fgpG(:,:,7), [0 1]); title('FGP-TV denoised volume (GPU)'); +%% +fprintf('Denoise a volume using the SB-TV model (CPU) \n'); +iter_sb = 150; % number of SB iterations +epsil_tol = 1.0e-05; % tolerance +tic; u_sb = SB_TV(single(vol3D), lambda_reg, iter_sb, epsil_tol); toc; +energyfunc_val_sb = TV_energy(single(u_sb),single(vol3D),lambda_reg, 1); % get energy function value +rmse_sb = (RMSE(Ideal3D(:),u_sb(:))); +fprintf('%s %f \n', 'RMSE error for SB-TV is:', rmse_sb); +figure; imshow(u_sb(:,:,7), [0 1]); title('SB-TV denoised volume (CPU)'); +%% +% fprintf('Denoise a volume using the SB-TV model (GPU) \n'); +% iter_sb = 150; % number of SB iterations +% epsil_tol = 1.0e-05; % tolerance +% tic; u_sbG = SB_TV_GPU(single(vol3D), lambda_reg, iter_sb, epsil_tol); toc; +% rmse_sbG = (RMSE(Ideal3D(:),u_sbG(:))); +% fprintf('%s %f \n', 'RMSE error for SB-TV is:', rmse_sbG); +% figure; imshow(u_sbG(:,:,7), [0 1]); title('SB-TV denoised volume (GPU)'); +%% +fprintf('Denoise a volume using the ROF-LLT model (CPU) \n'); +lambda_ROF = lambda_reg; % ROF regularisation parameter +lambda_LLT = lambda_reg*0.35; % LLT regularisation parameter +iter_LLT = 300; % iterations +tau_rof_llt = 0.0025; % time-marching constant +tic; u_rof_llt = LLT_ROF(single(vol3D), lambda_ROF, lambda_LLT, iter_LLT, tau_rof_llt); toc; +rmse_rof_llt = (RMSE(Ideal3D(:),u_rof_llt(:))); +fprintf('%s %f \n', 'RMSE error for ROF-LLT is:', rmse_rof_llt); +figure; imshow(u_rof_llt(:,:,7), [0 1]); title('ROF-LLT denoised volume (CPU)'); +%% +% fprintf('Denoise a volume using the ROF-LLT model (GPU) \n'); +% lambda_ROF = lambda_reg; % ROF regularisation parameter +% lambda_LLT = lambda_reg*0.35; % LLT regularisation parameter +% iter_LLT = 300; % iterations +% tau_rof_llt = 0.0025; % time-marching constant +% tic; u_rof_llt_g = LLT_ROF_GPU(single(vol3D), lambda_ROF, lambda_LLT, iter_LLT, tau_rof_llt); toc; +% rmse_rof_llt = (RMSE(Ideal3D(:),u_rof_llt_g(:))); +% fprintf('%s %f \n', 'RMSE error for ROF-LLT is:', rmse_rof_llt); +% figure; imshow(u_rof_llt_g(:,:,7), [0 1]); title('ROF-LLT denoised volume (GPU)'); +%% +fprintf('Denoise a volume using Nonlinear-Diffusion model (CPU) \n'); +iter_diff = 300; % number of diffusion iterations +lambda_regDiff = 0.025; % regularisation for the diffusivity +sigmaPar = 0.015; % edge-preserving parameter +tau_param = 0.025; % time-marching constant +tic; u_diff = NonlDiff(single(vol3D), lambda_regDiff, sigmaPar, iter_diff, tau_param, 'Huber'); toc; +rmse_diff = (RMSE(Ideal3D(:),u_diff(:))); +fprintf('%s %f \n', 'RMSE error for Diffusion is:', rmse_diff); +figure; imshow(u_diff(:,:,7), [0 1]); title('Diffusion denoised volume (CPU)'); +%% +% fprintf('Denoise a volume using Nonlinear-Diffusion model (GPU) \n'); +% iter_diff = 300; % number of diffusion iterations +% lambda_regDiff = 0.025; % regularisation for the diffusivity +% sigmaPar = 0.015; % edge-preserving parameter +% tau_param = 0.025; % time-marching constant +% tic; u_diff_g = NonlDiff_GPU(single(vol3D), lambda_regDiff, sigmaPar, iter_diff, tau_param, 'Huber'); toc; +% rmse_diff = (RMSE(Ideal3D(:),u_diff_g(:))); +% fprintf('%s %f \n', 'RMSE error for Diffusion is:', rmse_diff); +% figure; imshow(u_diff_g(:,:,7), [0 1]); title('Diffusion denoised volume (GPU)'); +%% +fprintf('Denoise using Fourth-order anisotropic diffusion model (CPU) \n'); +iter_diff = 300; % number of diffusion iterations +lambda_regDiff = 3.5; % regularisation for the diffusivity +sigmaPar = 0.02; % edge-preserving parameter +tau_param = 0.0015; % time-marching constant +tic; u_diff4 = Diffusion_4thO(single(vol3D), lambda_regDiff, sigmaPar, iter_diff, tau_param); toc; +rmse_diff4 = (RMSE(Ideal3D(:),u_diff4(:))); +fprintf('%s %f \n', 'RMSE error for Anis.Diff of 4th order is:', rmse_diff4); +figure; imshow(u_diff4(:,:,7), [0 1]); title('Diffusion 4thO denoised volume (CPU)'); +%% +% fprintf('Denoise using Fourth-order anisotropic diffusion model (GPU) \n'); +% iter_diff = 300; % number of diffusion iterations +% lambda_regDiff = 3.5; % regularisation for the diffusivity +% sigmaPar = 0.02; % edge-preserving parameter +% tau_param = 0.0015; % time-marching constant +% tic; u_diff4_g = Diffusion_4thO_GPU(single(vol3D), lambda_regDiff, sigmaPar, iter_diff, tau_param); toc; +% rmse_diff4 = (RMSE(Ideal3D(:),u_diff4_g(:))); +% fprintf('%s %f \n', 'RMSE error for Anis.Diff of 4th order is:', rmse_diff4); +% figure; imshow(u_diff4_g(:,:,7), [0 1]); title('Diffusion 4thO denoised volume (GPU)'); +%% +fprintf('Denoise using the TGV model (CPU) \n'); +lambda_TGV = 0.03; % regularisation parameter +alpha1 = 1.0; % parameter to control the first-order term +alpha0 = 2.0; % parameter to control the second-order term +iter_TGV = 500; % number of Primal-Dual iterations for TGV +tic; u_tgv = TGV(single(vol3D), lambda_TGV, alpha1, alpha0, iter_TGV); toc; +rmseTGV = RMSE(Ideal3D(:),u_tgv(:)); +fprintf('%s %f \n', 'RMSE error for TGV is:', rmseTGV); +figure; imshow(u_tgv(:,:,3), [0 1]); title('TGV denoised volume (CPU)'); +%% +%>>>>>>>>>>>>>> MULTI-CHANNEL priors <<<<<<<<<<<<<<< % +fprintf('Denoise a volume using the FGP-dTV model (CPU) \n'); + +% create another volume (reference) with slightly less amount of noise +vol3D_ref = zeros(N,N,slices, 'single'); +for i = 1:slices +vol3D_ref(:,:,i) = Im + .01*randn(size(Im)); +end +vol3D_ref(vol3D_ref < 0) = 0; +% vol3D_ref = zeros(size(Im),'single'); % pass zero reference (dTV -> TV) + +iter_fgp = 300; % number of FGP iterations +epsil_tol = 1.0e-05; % tolerance +eta = 0.2; % Reference image gradient smoothing constant +tic; u_fgp_dtv = FGP_dTV(single(vol3D), single(vol3D_ref), lambda_reg, iter_fgp, epsil_tol, eta); toc; +figure; imshow(u_fgp_dtv(:,:,7), [0 1]); title('FGP-dTV denoised volume (CPU)'); +%% +fprintf('Denoise a volume using the FGP-dTV model (GPU) \n'); + +% create another volume (reference) with slightly less amount of noise +vol3D_ref = zeros(N,N,slices, 'single'); +for i = 1:slices +vol3D_ref(:,:,i) = Im + .01*randn(size(Im)); +end +vol3D_ref(vol3D_ref < 0) = 0; +% vol3D_ref = zeros(size(Im),'single'); % pass zero reference (dTV -> TV) + +iter_fgp = 300; % number of FGP iterations +epsil_tol = 1.0e-05; % tolerance +eta = 0.2; % Reference image gradient smoothing constant +tic; u_fgp_dtv_g = FGP_dTV_GPU(single(vol3D), single(vol3D_ref), lambda_reg, iter_fgp, epsil_tol, eta); toc; +figure; imshow(u_fgp_dtv_g(:,:,7), [0 1]); title('FGP-dTV denoised volume (GPU)'); +%% diff --git a/docs/demos/demoMatlab_denoise.m b/docs/demos/demoMatlab_denoise.m new file mode 100644 index 0000000..14d3096 --- /dev/null +++ b/docs/demos/demoMatlab_denoise.m @@ -0,0 +1,189 @@ +% Image (2D) denoising demo using CCPi-RGL +clear; close all +fsep = '/'; + +Path1 = sprintf(['..' fsep 'mex_compile' fsep 'installed'], 1i); +Path2 = sprintf(['..' fsep '..' fsep '..' fsep 'data' fsep], 1i); +Path3 = sprintf(['..' fsep 'supp'], 1i); +addpath(Path1); addpath(Path2); addpath(Path3); + +Im = double(imread('lena_gray_512.tif'))/255; % loading image +u0 = Im + .05*randn(size(Im)); u0(u0 < 0) = 0; +figure; imshow(u0, [0 1]); title('Noisy image'); + +lambda_reg = 0.03; % regularsation parameter for all methods +%% +fprintf('Denoise using the ROF-TV model (CPU) \n'); +tau_rof = 0.0025; % time-marching constant +iter_rof = 750; % number of ROF iterations +tic; u_rof = ROF_TV(single(u0), lambda_reg, iter_rof, tau_rof); toc; +energyfunc_val_rof = TV_energy(single(u_rof),single(u0),lambda_reg, 1); % get energy function value +rmseROF = (RMSE(u_rof(:),Im(:))); +fprintf('%s %f \n', 'RMSE error for ROF-TV is:', rmseROF); +figure; imshow(u_rof, [0 1]); title('ROF-TV denoised image (CPU)'); +%% +% fprintf('Denoise using the ROF-TV model (GPU) \n'); +% tau_rof = 0.0025; % time-marching constant +% iter_rof = 750; % number of ROF iterations +% tic; u_rofG = ROF_TV_GPU(single(u0), lambda_reg, iter_rof, tau_rof); toc; +% figure; imshow(u_rofG, [0 1]); title('ROF-TV denoised image (GPU)'); +%% +fprintf('Denoise using the FGP-TV model (CPU) \n'); +iter_fgp = 1000; % number of FGP iterations +epsil_tol = 1.0e-06; % tolerance +tic; u_fgp = FGP_TV(single(u0), lambda_reg, iter_fgp, epsil_tol); toc; +energyfunc_val_fgp = TV_energy(single(u_fgp),single(u0),lambda_reg, 1); % get energy function value +rmseFGP = (RMSE(u_fgp(:),Im(:))); +fprintf('%s %f \n', 'RMSE error for FGP-TV is:', rmseFGP); +figure; imshow(u_fgp, [0 1]); title('FGP-TV denoised image (CPU)'); + +%% +% fprintf('Denoise using the FGP-TV model (GPU) \n'); +% iter_fgp = 1000; % number of FGP iterations +% epsil_tol = 1.0e-05; % tolerance +% tic; u_fgpG = FGP_TV_GPU(single(u0), lambda_reg, iter_fgp, epsil_tol); toc; +% figure; imshow(u_fgpG, [0 1]); title('FGP-TV denoised image (GPU)'); +%% +fprintf('Denoise using the SB-TV model (CPU) \n'); +iter_sb = 150; % number of SB iterations +epsil_tol = 1.0e-06; % tolerance +tic; u_sb = SB_TV(single(u0), lambda_reg, iter_sb, epsil_tol); toc; +energyfunc_val_sb = TV_energy(single(u_sb),single(u0),lambda_reg, 1); % get energy function value +rmseSB = (RMSE(u_sb(:),Im(:))); +fprintf('%s %f \n', 'RMSE error for SB-TV is:', rmseSB); +figure; imshow(u_sb, [0 1]); title('SB-TV denoised image (CPU)'); +%% +% fprintf('Denoise using the SB-TV model (GPU) \n'); +% iter_sb = 150; % number of SB iterations +% epsil_tol = 1.0e-06; % tolerance +% tic; u_sbG = SB_TV_GPU(single(u0), lambda_reg, iter_sb, epsil_tol); toc; +% figure; imshow(u_sbG, [0 1]); title('SB-TV denoised image (GPU)'); +%% +fprintf('Denoise using the TGV model (CPU) \n'); +lambda_TGV = 0.045; % regularisation parameter +alpha1 = 1.0; % parameter to control the first-order term +alpha0 = 2.0; % parameter to control the second-order term +iter_TGV = 2000; % number of Primal-Dual iterations for TGV +tic; u_tgv = TGV(single(u0), lambda_TGV, alpha1, alpha0, iter_TGV); toc; +rmseTGV = (RMSE(u_tgv(:),Im(:))); +fprintf('%s %f \n', 'RMSE error for TGV is:', rmseTGV); +figure; imshow(u_tgv, [0 1]); title('TGV denoised image (CPU)'); +%% +% fprintf('Denoise using the TGV model (GPU) \n'); +% lambda_TGV = 0.045; % regularisation parameter +% alpha1 = 1.0; % parameter to control the first-order term +% alpha0 = 2.0; % parameter to control the second-order term +% iter_TGV = 2000; % number of Primal-Dual iterations for TGV +% tic; u_tgv_gpu = TGV_GPU(single(u0), lambda_TGV, alpha1, alpha0, iter_TGV); toc; +% rmseTGV_gpu = (RMSE(u_tgv_gpu(:),Im(:))); +% fprintf('%s %f \n', 'RMSE error for TGV is:', rmseTGV_gpu); +% figure; imshow(u_tgv_gpu, [0 1]); title('TGV denoised image (GPU)'); +%% +fprintf('Denoise using the ROF-LLT model (CPU) \n'); +lambda_ROF = lambda_reg; % ROF regularisation parameter +lambda_LLT = lambda_reg*0.45; % LLT regularisation parameter +iter_LLT = 1; % iterations +tau_rof_llt = 0.0025; % time-marching constant +tic; u_rof_llt = LLT_ROF(single(u0), lambda_ROF, lambda_LLT, iter_LLT, tau_rof_llt); toc; +rmseROFLLT = (RMSE(u_rof_llt(:),Im(:))); +fprintf('%s %f \n', 'RMSE error for TGV is:', rmseROFLLT); +figure; imshow(u_rof_llt, [0 1]); title('ROF-LLT denoised image (CPU)'); +%% +% fprintf('Denoise using the ROF-LLT model (GPU) \n'); +% lambda_ROF = lambda_reg; % ROF regularisation parameter +% lambda_LLT = lambda_reg*0.45; % LLT regularisation parameter +% iter_LLT = 500; % iterations +% tau_rof_llt = 0.0025; % time-marching constant +% tic; u_rof_llt_g = LLT_ROF_GPU(single(u0), lambda_ROF, lambda_LLT, iter_LLT, tau_rof_llt); toc; +% rmseROFLLT_g = (RMSE(u_rof_llt_g(:),Im(:))); +% fprintf('%s %f \n', 'RMSE error for TGV is:', rmseROFLLT_g); +% figure; imshow(u_rof_llt_g, [0 1]); title('ROF-LLT denoised image (GPU)'); +%% +fprintf('Denoise using Nonlinear-Diffusion model (CPU) \n'); +iter_diff = 800; % number of diffusion iterations +lambda_regDiff = 0.025; % regularisation for the diffusivity +sigmaPar = 0.015; % edge-preserving parameter +tau_param = 0.025; % time-marching constant +tic; u_diff = NonlDiff(single(u0), lambda_regDiff, sigmaPar, iter_diff, tau_param, 'Huber'); toc; +rmseDiffus = (RMSE(u_diff(:),Im(:))); +fprintf('%s %f \n', 'RMSE error for Nonlinear Diffusion is:', rmseDiffus); +figure; imshow(u_diff, [0 1]); title('Diffusion denoised image (CPU)'); +%% +% fprintf('Denoise using Nonlinear-Diffusion model (GPU) \n'); +% iter_diff = 800; % number of diffusion iterations +% lambda_regDiff = 0.025; % regularisation for the diffusivity +% sigmaPar = 0.015; % edge-preserving parameter +% tau_param = 0.025; % time-marching constant +% tic; u_diff_g = NonlDiff_GPU(single(u0), lambda_regDiff, sigmaPar, iter_diff, tau_param, 'Huber'); toc; +% figure; imshow(u_diff_g, [0 1]); title('Diffusion denoised image (GPU)'); +%% +fprintf('Denoise using Fourth-order anisotropic diffusion model (CPU) \n'); +iter_diff = 800; % number of diffusion iterations +lambda_regDiff = 3.5; % regularisation for the diffusivity +sigmaPar = 0.02; % edge-preserving parameter +tau_param = 0.0015; % time-marching constant +tic; u_diff4 = Diffusion_4thO(single(u0), lambda_regDiff, sigmaPar, iter_diff, tau_param); toc; +rmseDiffHO = (RMSE(u_diff4(:),Im(:))); +fprintf('%s %f \n', 'RMSE error for Fourth-order anisotropic diffusion is:', rmseDiffHO); +figure; imshow(u_diff4, [0 1]); title('Diffusion 4thO denoised image (CPU)'); +%% +% fprintf('Denoise using Fourth-order anisotropic diffusion model (GPU) \n'); +% iter_diff = 800; % number of diffusion iterations +% lambda_regDiff = 3.5; % regularisation for the diffusivity +% sigmaPar = 0.02; % edge-preserving parameter +% tau_param = 0.0015; % time-marching constant +% tic; u_diff4_g = Diffusion_4thO_GPU(single(u0), lambda_regDiff, sigmaPar, iter_diff, tau_param); toc; +% figure; imshow(u_diff4_g, [0 1]); title('Diffusion 4thO denoised image (GPU)'); +%% +fprintf('Weights pre-calculation for Non-local TV (takes time on CPU) \n'); +SearchingWindow = 7; +PatchWindow = 2; +NeighboursNumber = 20; % the number of neibours to include +h = 0.23; % edge related parameter for NLM +tic; [H_i, H_j, Weights] = PatchSelect(single(u0), SearchingWindow, PatchWindow, NeighboursNumber, h); toc; +%% +fprintf('Denoise using Non-local Total Variation (CPU) \n'); +iter_nltv = 3; % number of nltv iterations +lambda_nltv = 0.05; % regularisation parameter for nltv +tic; u_nltv = Nonlocal_TV(single(u0), H_i, H_j, 0, Weights, lambda_nltv, iter_nltv); toc; +rmse_nltv = (RMSE(u_nltv(:),Im(:))); +fprintf('%s %f \n', 'RMSE error for Non-local Total Variation is:', rmse_nltv); +figure; imagesc(u_nltv, [0 1]); colormap(gray); daspect([1 1 1]); title('Non-local Total Variation denoised image (CPU)'); +%% +%>>>>>>>>>>>>>> MULTI-CHANNEL priors <<<<<<<<<<<<<<< % + +fprintf('Denoise using the FGP-dTV model (CPU) \n'); +% create another image (reference) with slightly less amount of noise +u_ref = Im + .01*randn(size(Im)); u_ref(u_ref < 0) = 0; +% u_ref = zeros(size(Im),'single'); % pass zero reference (dTV -> TV) + +iter_fgp = 1000; % number of FGP iterations +epsil_tol = 1.0e-06; % tolerance +eta = 0.2; % Reference image gradient smoothing constant +tic; u_fgp_dtv = FGP_dTV(single(u0), single(u_ref), lambda_reg, iter_fgp, epsil_tol, eta); toc; +rmse_dTV= (RMSE(u_fgp_dtv(:),Im(:))); +fprintf('%s %f \n', 'RMSE error for Directional Total Variation (dTV) is:', rmse_dTV); +figure; imshow(u_fgp_dtv, [0 1]); title('FGP-dTV denoised image (CPU)'); +%% +% fprintf('Denoise using the FGP-dTV model (GPU) \n'); +% % create another image (reference) with slightly less amount of noise +% u_ref = Im + .01*randn(size(Im)); u_ref(u_ref < 0) = 0; +% % u_ref = zeros(size(Im),'single'); % pass zero reference (dTV -> TV) +% +% iter_fgp = 1000; % number of FGP iterations +% epsil_tol = 1.0e-06; % tolerance +% eta = 0.2; % Reference image gradient smoothing constant +% tic; u_fgp_dtvG = FGP_dTV_GPU(single(u0), single(u_ref), lambda_reg, iter_fgp, epsil_tol, eta); toc; +% figure; imshow(u_fgp_dtvG, [0 1]); title('FGP-dTV denoised image (GPU)'); +%% +fprintf('Denoise using the TNV prior (CPU) \n'); +slices = 5; N = 512; +vol3D = zeros(N,N,slices, 'single'); +for i = 1:slices +vol3D(:,:,i) = Im + .05*randn(size(Im)); +end +vol3D(vol3D < 0) = 0; + +iter_tnv = 200; % number of TNV iterations +tic; u_tnv = TNV(single(vol3D), lambda_reg, iter_tnv); toc; +figure; imshow(u_tnv(:,:,3), [0 1]); title('TNV denoised stack of channels (CPU)'); diff --git a/docs/demos/demoMatlab_inpaint.m b/docs/demos/demoMatlab_inpaint.m new file mode 100644 index 0000000..66f9c15 --- /dev/null +++ b/docs/demos/demoMatlab_inpaint.m @@ -0,0 +1,35 @@ +% Image (2D) inpainting demo using CCPi-RGL +clear; close all +Path1 = sprintf(['..' filesep 'mex_compile' filesep 'installed'], 1i); +Path2 = sprintf(['..' filesep '..' filesep '..' filesep 'data' filesep], 1i); +addpath(Path1); +addpath(Path2); + +load('SinoInpaint.mat'); +Sinogram = Sinogram./max(Sinogram(:)); +Sino_mask = Sinogram.*(1-single(Mask)); +figure; +subplot(1,2,1); imshow(Sino_mask, [0 1]); title('Missing data sinogram'); +subplot(1,2,2); imshow(Mask, [0 1]); title('Mask'); +%% +fprintf('Inpaint using Linear-Diffusion model (CPU) \n'); +iter_diff = 5000; % number of diffusion iterations +lambda_regDiff = 6000; % regularisation for the diffusivity +sigmaPar = 0.0; % edge-preserving parameter +tau_param = 0.000075; % time-marching constant +tic; u_diff = NonlDiff_Inp(single(Sino_mask), Mask, lambda_regDiff, sigmaPar, iter_diff, tau_param); toc; +figure; imshow(u_diff, [0 1]); title('Linear-Diffusion inpainted sinogram (CPU)'); +%% +fprintf('Inpaint using Nonlinear-Diffusion model (CPU) \n'); +iter_diff = 1500; % number of diffusion iterations +lambda_regDiff = 80; % regularisation for the diffusivity +sigmaPar = 0.00009; % edge-preserving parameter +tau_param = 0.000008; % time-marching constant +tic; u_diff = NonlDiff_Inp(single(Sino_mask), Mask, lambda_regDiff, sigmaPar, iter_diff, tau_param, 'Huber'); toc; +figure; imshow(u_diff, [0 1]); title('Non-Linear Diffusion inpainted sinogram (CPU)'); +%% +fprintf('Inpaint using Nonlocal Vertical Marching model (CPU) \n'); +Increment = 1; % linear increment for the searching window +tic; [u_nom,maskupd] = NonlocalMarching_Inpaint(single(Sino_mask), Mask, Increment); toc; +figure; imshow(u_nom, [0 1]); title('NVM inpainted sinogram (CPU)'); +%%
\ No newline at end of file diff --git a/docs/demos/demo_cpu_inpainters.py b/docs/demos/demo_cpu_inpainters.py new file mode 100644 index 0000000..3b4191b --- /dev/null +++ b/docs/demos/demo_cpu_inpainters.py @@ -0,0 +1,192 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Demonstration of CPU inpainters +@authors: Daniil Kazantsev, Edoardo Pasca +""" + +import matplotlib.pyplot as plt +import numpy as np +import os +import timeit +from scipy import io +from ccpi.filters.regularisers import NDF_INP, NVM_INP +from qualitymetrics import rmse +############################################################################### +def printParametersToString(pars): + txt = r'' + for key, value in pars.items(): + if key== 'algorithm' : + txt += "{0} = {1}".format(key, value.__name__) + elif key == 'input': + txt += "{0} = {1}".format(key, np.shape(value)) + elif key == 'maskData': + txt += "{0} = {1}".format(key, np.shape(value)) + else: + txt += "{0} = {1}".format(key, value) + txt += '\n' + return txt +############################################################################### + +# read sinogram and the mask +filename = os.path.join(".." , ".." , ".." , "data" ,"SinoInpaint.mat") +sino = io.loadmat(filename) +sino_full = sino.get('Sinogram') +Mask = sino.get('Mask') +[angles_dim,detectors_dim] = sino_full.shape +sino_full = sino_full/np.max(sino_full) +#apply mask to sinogram +sino_cut = sino_full*(1-Mask) +#sino_cut_new = np.zeros((angles_dim,detectors_dim),'float32') +#sino_cut_new = sino_cut.copy(order='c') +#sino_cut_new[:] = sino_cut[:] +sino_cut_new = np.ascontiguousarray(sino_cut, dtype=np.float32); +#mask = np.zeros((angles_dim,detectors_dim),'uint8') +#mask =Mask.copy(order='c') +#mask[:] = Mask[:] +mask = np.ascontiguousarray(Mask, dtype=np.uint8); + +plt.figure(1) +plt.subplot(121) +plt.imshow(sino_cut_new,vmin=0.0, vmax=1) +plt.title('Missing Data sinogram') +plt.subplot(122) +plt.imshow(mask) +plt.title('Mask') +plt.show() +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Inpainting using linear diffusion (2D)__") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure(2) +plt.suptitle('Performance of linear inpainting using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Missing data sinogram') +imgplot = plt.imshow(sino_cut_new,cmap="gray") + +# set parameters +pars = {'algorithm' : NDF_INP, \ + 'input' : sino_cut_new,\ + 'maskData' : mask,\ + 'regularisation_parameter':5000,\ + 'edge_parameter':0,\ + 'number_of_iterations' :5000 ,\ + 'time_marching_parameter':0.000075,\ + 'penalty_type':0 + } + +start_time = timeit.default_timer() +ndf_inp_linear = NDF_INP(pars['input'], + pars['maskData'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], + pars['penalty_type']) + +rms = rmse(sino_full, ndf_inp_linear) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(ndf_inp_linear, cmap="gray") +plt.title('{}'.format('Linear diffusion inpainting results')) +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_Inpainting using nonlinear diffusion (2D)_") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure(3) +plt.suptitle('Performance of nonlinear diffusion inpainting using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Missing data sinogram') +imgplot = plt.imshow(sino_cut_new,cmap="gray") + +# set parameters +pars = {'algorithm' : NDF_INP, \ + 'input' : sino_cut_new,\ + 'maskData' : mask,\ + 'regularisation_parameter':80,\ + 'edge_parameter':0.00009,\ + 'number_of_iterations' :1500 ,\ + 'time_marching_parameter':0.000008,\ + 'penalty_type':1 + } + +start_time = timeit.default_timer() +ndf_inp_nonlinear = NDF_INP(pars['input'], + pars['maskData'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], + pars['penalty_type']) + +rms = rmse(sino_full, ndf_inp_nonlinear) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(ndf_inp_nonlinear, cmap="gray") +plt.title('{}'.format('Nonlinear diffusion inpainting results')) +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("Inpainting using nonlocal vertical marching") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure(4) +plt.suptitle('Performance of NVM inpainting using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Missing data sinogram') +imgplot = plt.imshow(sino_cut,cmap="gray") + +# set parameters +pars = {'algorithm' : NVM_INP, \ + 'input' : sino_cut_new,\ + 'maskData' : mask,\ + 'SW_increment': 1,\ + 'number_of_iterations' : 150 + } + +start_time = timeit.default_timer() +(nvm_inp, mask_upd) = NVM_INP(pars['input'], + pars['maskData'], + pars['SW_increment'], + pars['number_of_iterations']) + +rms = rmse(sino_full, nvm_inp) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(nvm_inp, cmap="gray") +plt.title('{}'.format('Nonlocal Vertical Marching inpainting results')) +#%% diff --git a/docs/demos/demo_cpu_regularisers.py b/docs/demos/demo_cpu_regularisers.py new file mode 100644 index 0000000..e6befa9 --- /dev/null +++ b/docs/demos/demo_cpu_regularisers.py @@ -0,0 +1,572 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Thu Feb 22 11:39:43 2018 + +Demonstration of CPU regularisers + +@authors: Daniil Kazantsev, Edoardo Pasca +""" + +import matplotlib.pyplot as plt +import numpy as np +import os +import timeit +from ccpi.filters.regularisers import ROF_TV, FGP_TV, SB_TV, TGV, LLT_ROF, FGP_dTV, TNV, NDF, Diff4th +from ccpi.filters.regularisers import PatchSelect, NLTV +from qualitymetrics import rmse +############################################################################### +def printParametersToString(pars): + txt = r'' + for key, value in pars.items(): + if key== 'algorithm' : + txt += "{0} = {1}".format(key, value.__name__) + elif key == 'input': + txt += "{0} = {1}".format(key, np.shape(value)) + elif key == 'refdata': + txt += "{0} = {1}".format(key, np.shape(value)) + else: + txt += "{0} = {1}".format(key, value) + txt += '\n' + return txt +############################################################################### +#%% +filename = os.path.join(".." , ".." , ".." , "data" ,"lena_gray_512.tif") + +# read image +Im = plt.imread(filename) +Im = np.asarray(Im, dtype='float32') + +Im = Im/255.0 +perc = 0.05 +u0 = Im + np.random.normal(loc = 0 , + scale = perc * Im , + size = np.shape(Im)) +u_ref = Im + np.random.normal(loc = 0 , + scale = 0.01 * Im , + size = np.shape(Im)) +(N,M) = np.shape(u0) +# map the u0 u0->u0>0 +# f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1) +u0 = u0.astype('float32') +u_ref = u_ref.astype('float32') + +# change dims to check that modules work with non-squared images +""" +M = M-100 +u_ref2 = np.zeros([N,M],dtype='float32') +u_ref2[:,0:M] = u_ref[:,0:M] +u_ref = u_ref2 +del u_ref2 + +u02 = np.zeros([N,M],dtype='float32') +u02[:,0:M] = u0[:,0:M] +u0 = u02 +del u02 + +Im2 = np.zeros([N,M],dtype='float32') +Im2[:,0:M] = Im[:,0:M] +Im = Im2 +del Im2 +""" +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________ROF-TV (2D)_________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of ROF-TV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm': ROF_TV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04,\ + 'number_of_iterations': 1200,\ + 'time_marching_parameter': 0.0025 + } +print ("#############ROF TV CPU####################") +start_time = timeit.default_timer() +rof_cpu = ROF_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'cpu') +rms = rmse(Im, rof_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(rof_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________FGP-TV (2D)__________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of FGP-TV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_TV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :2000 ,\ + 'tolerance_constant':1e-06,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("#############FGP TV CPU####################") +start_time = timeit.default_timer() +fgp_cpu = FGP_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'cpu') + + +rms = rmse(Im, fgp_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________SB-TV (2D)__________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of SB-TV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : SB_TV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :150 ,\ + 'tolerance_constant':1e-06,\ + 'methodTV': 0 ,\ + 'printingOut': 0 + } + +print ("#############SB TV CPU####################") +start_time = timeit.default_timer() +sb_cpu = SB_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['printingOut'],'cpu') + + +rms = rmse(Im, sb_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(sb_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) +#%% + +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_____Total Generalised Variation (2D)______") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of TGV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : TGV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04, \ + 'alpha1':1.0,\ + 'alpha0':2.0,\ + 'number_of_iterations' :1350 ,\ + 'LipshitzConstant' :12 ,\ + } + +print ("#############TGV CPU####################") +start_time = timeit.default_timer() +tgv_cpu = TGV(pars['input'], + pars['regularisation_parameter'], + pars['alpha1'], + pars['alpha0'], + pars['number_of_iterations'], + pars['LipshitzConstant'],'cpu') + + +rms = rmse(Im, tgv_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(tgv_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +#%% + +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("______________LLT- ROF (2D)________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of LLT-ROF regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : LLT_ROF, \ + 'input' : u0,\ + 'regularisation_parameterROF':0.04, \ + 'regularisation_parameterLLT':0.01, \ + 'number_of_iterations' :500 ,\ + 'time_marching_parameter' :0.0025 ,\ + } + +print ("#############LLT- ROF CPU####################") +start_time = timeit.default_timer() +lltrof_cpu = LLT_ROF(pars['input'], + pars['regularisation_parameterROF'], + pars['regularisation_parameterLLT'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'cpu') + +rms = rmse(Im, lltrof_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(lltrof_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +#%% + + +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("________________NDF (2D)___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of NDF regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : NDF, \ + 'input' : u0,\ + 'regularisation_parameter':0.025, \ + 'edge_parameter':0.015,\ + 'number_of_iterations' :500 ,\ + 'time_marching_parameter':0.025,\ + 'penalty_type':1 + } + +print ("#############NDF CPU################") +start_time = timeit.default_timer() +ndf_cpu = NDF(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], + pars['penalty_type'],'cpu') + +rms = rmse(Im, ndf_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(ndf_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Anisotropic Diffusion 4th Order (2D)____") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of Diff4th regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : Diff4th, \ + 'input' : u0,\ + 'regularisation_parameter':3.5, \ + 'edge_parameter':0.02,\ + 'number_of_iterations' :500 ,\ + 'time_marching_parameter':0.0015 + } + +print ("#############Diff4th CPU################") +start_time = timeit.default_timer() +diff4_cpu = Diff4th(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'cpu') + +rms = rmse(Im, diff4_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(diff4_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Nonlocal patches pre-calculation____") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +start_time = timeit.default_timer() +# set parameters +pars = {'algorithm' : PatchSelect, \ + 'input' : u0,\ + 'searchwindow': 7, \ + 'patchwindow': 2,\ + 'neighbours' : 15 ,\ + 'edge_parameter':0.18} + +H_i, H_j, Weights = PatchSelect(pars['input'], + pars['searchwindow'], + pars['patchwindow'], + pars['neighbours'], + pars['edge_parameter'],'cpu') + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +""" +plt.figure() +plt.imshow(Weights[0,:,:],cmap="gray",interpolation="nearest",vmin=0, vmax=1) +plt.show() +""" +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Nonlocal Total Variation penalty____") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +## plot +fig = plt.figure() +plt.suptitle('Performance of NLTV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +pars2 = {'algorithm' : NLTV, \ + 'input' : u0,\ + 'H_i': H_i, \ + 'H_j': H_j,\ + 'H_k' : 0,\ + 'Weights' : Weights,\ + 'regularisation_parameter': 0.04,\ + 'iterations': 3 + } +start_time = timeit.default_timer() +nltv_cpu = NLTV(pars2['input'], + pars2['H_i'], + pars2['H_j'], + pars2['H_k'], + pars2['Weights'], + pars2['regularisation_parameter'], + pars2['iterations']) + +rms = rmse(Im, nltv_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(nltv_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_____________FGP-dTV (2D)__________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of FGP-dTV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_dTV, \ + 'input' : u0,\ + 'refdata' : u_ref,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :2000 ,\ + 'tolerance_constant':1e-06,\ + 'eta_const':0.2,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("#############FGP dTV CPU####################") +start_time = timeit.default_timer() +fgp_dtv_cpu = FGP_dTV(pars['input'], + pars['refdata'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['eta_const'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'cpu') + +rms = rmse(Im, fgp_dtv_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_dtv_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("__________Total nuclear Variation__________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of TNV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +channelsNo = 5 +noisyVol = np.zeros((channelsNo,N,M),dtype='float32') +idealVol = np.zeros((channelsNo,N,M),dtype='float32') + +for i in range (channelsNo): + noisyVol[i,:,:] = Im + np.random.normal(loc = 0 , scale = perc * Im , size = np.shape(Im)) + idealVol[i,:,:] = Im + +# set parameters +pars = {'algorithm' : TNV, \ + 'input' : noisyVol,\ + 'regularisation_parameter': 0.04, \ + 'number_of_iterations' : 200 ,\ + 'tolerance_constant':1e-05 + } + +print ("#############TNV CPU#################") +start_time = timeit.default_timer() +tnv_cpu = TNV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant']) + +rms = rmse(idealVol, tnv_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(tnv_cpu[3,:,:], cmap="gray") +plt.title('{}'.format('CPU results')) diff --git a/docs/demos/demo_cpu_regularisers3D.py b/docs/demos/demo_cpu_regularisers3D.py new file mode 100644 index 0000000..2d2fc22 --- /dev/null +++ b/docs/demos/demo_cpu_regularisers3D.py @@ -0,0 +1,458 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Thu Feb 22 11:39:43 2018 + +Demonstration of 3D CPU regularisers + +@authors: Daniil Kazantsev, Edoardo Pasca +""" + +import matplotlib.pyplot as plt +import numpy as np +import os +import timeit +from ccpi.filters.regularisers import ROF_TV, FGP_TV, SB_TV, TGV, LLT_ROF, FGP_dTV, NDF, Diff4th +from qualitymetrics import rmse +############################################################################### +def printParametersToString(pars): + txt = r'' + for key, value in pars.items(): + if key== 'algorithm' : + txt += "{0} = {1}".format(key, value.__name__) + elif key == 'input': + txt += "{0} = {1}".format(key, np.shape(value)) + elif key == 'refdata': + txt += "{0} = {1}".format(key, np.shape(value)) + else: + txt += "{0} = {1}".format(key, value) + txt += '\n' + return txt +############################################################################### +#%% +filename = os.path.join(".." , ".." , ".." , "data" ,"lena_gray_512.tif") + +# read image +Im = plt.imread(filename) +Im = np.asarray(Im, dtype='float32') + +Im = Im/255 +perc = 0.05 +u0 = Im + np.random.normal(loc = 0 , + scale = perc * Im , + size = np.shape(Im)) +u_ref = Im + np.random.normal(loc = 0 , + scale = 0.01 * Im , + size = np.shape(Im)) +(N,M) = np.shape(u0) +# map the u0 u0->u0>0 +# f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1) +u0 = u0.astype('float32') +u_ref = u_ref.astype('float32') + +# change dims to check that modules work with non-squared images +""" +M = M-100 +u_ref2 = np.zeros([N,M],dtype='float32') +u_ref2[:,0:M] = u_ref[:,0:M] +u_ref = u_ref2 +del u_ref2 + +u02 = np.zeros([N,M],dtype='float32') +u02[:,0:M] = u0[:,0:M] +u0 = u02 +del u02 + +Im2 = np.zeros([N,M],dtype='float32') +Im2[:,0:M] = Im[:,0:M] +Im = Im2 +del Im2 +""" +slices = 15 + +noisyVol = np.zeros((slices,N,M),dtype='float32') +noisyRef = np.zeros((slices,N,M),dtype='float32') +idealVol = np.zeros((slices,N,M),dtype='float32') + +for i in range (slices): + noisyVol[i,:,:] = Im + np.random.normal(loc = 0 , scale = perc * Im , size = np.shape(Im)) + noisyRef[i,:,:] = Im + np.random.normal(loc = 0 , scale = 0.01 * Im , size = np.shape(Im)) + idealVol[i,:,:] = Im + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________ROF-TV (3D)_________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of ROF-TV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy 15th slice of a volume') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm': ROF_TV, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.04,\ + 'number_of_iterations': 500,\ + 'time_marching_parameter': 0.0025 + } +print ("#############ROF TV CPU####################") +start_time = timeit.default_timer() +rof_cpu3D = ROF_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'cpu') +rms = rmse(idealVol, rof_cpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(rof_cpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the CPU using ROF-TV')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________FGP-TV (3D)__________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of FGP-TV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_TV, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :300 ,\ + 'tolerance_constant':0.00001,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("#############FGP TV CPU####################") +start_time = timeit.default_timer() +fgp_cpu3D = FGP_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'cpu') + + +rms = rmse(idealVol, fgp_cpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_cpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the CPU using FGP-TV')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________SB-TV (3D)_________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of SB-TV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : SB_TV, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :150 ,\ + 'tolerance_constant':0.00001,\ + 'methodTV': 0 ,\ + 'printingOut': 0 + } + +print ("#############SB TV CPU####################") +start_time = timeit.default_timer() +sb_cpu3D = SB_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['printingOut'],'cpu') + +rms = rmse(idealVol, sb_cpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(sb_cpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the CPU using SB-TV')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________LLT-ROF (3D)_________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of LLT-ROF regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : LLT_ROF, \ + 'input' : noisyVol,\ + 'regularisation_parameterROF':0.04, \ + 'regularisation_parameterLLT':0.015, \ + 'number_of_iterations' :300 ,\ + 'time_marching_parameter' :0.0025 ,\ + } + +print ("#############LLT ROF CPU####################") +start_time = timeit.default_timer() +lltrof_cpu3D = LLT_ROF(pars['input'], + pars['regularisation_parameterROF'], + pars['regularisation_parameterLLT'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'cpu') + +rms = rmse(idealVol, lltrof_cpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(lltrof_cpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the CPU using LLT-ROF')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________TGV (3D)_________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of TGV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : TGV, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.04, \ + 'alpha1':1.0,\ + 'alpha0':2.0,\ + 'number_of_iterations' :250 ,\ + 'LipshitzConstant' :12 ,\ + } + +print ("#############TGV CPU####################") +start_time = timeit.default_timer() +tgv_cpu3D = TGV(pars['input'], + pars['regularisation_parameter'], + pars['alpha1'], + pars['alpha0'], + pars['number_of_iterations'], + pars['LipshitzConstant'],'cpu') + + +rms = rmse(idealVol, tgv_cpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(tgv_cpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the CPU using TGV')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("________________NDF (3D)___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of NDF regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy volume') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : NDF, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.025, \ + 'edge_parameter':0.015,\ + 'number_of_iterations' :500 ,\ + 'time_marching_parameter':0.025,\ + 'penalty_type': 1 + } + +print ("#############NDF CPU################") +start_time = timeit.default_timer() +ndf_cpu3D = NDF(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], + pars['penalty_type']) + +rms = rmse(idealVol, ndf_cpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(ndf_cpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the CPU using NDF iterations')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Anisotropic Diffusion 4th Order (2D)____") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of Diff4th regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy volume') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : Diff4th, \ + 'input' : noisyVol,\ + 'regularisation_parameter':3.5, \ + 'edge_parameter':0.02,\ + 'number_of_iterations' :300 ,\ + 'time_marching_parameter':0.0015 + } + +print ("#############Diff4th CPU################") +start_time = timeit.default_timer() +diff4th_cpu3D = Diff4th(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter']) + +rms = rmse(idealVol, diff4th_cpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(diff4th_cpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the CPU using DIFF4th iterations')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________FGP-dTV (3D)__________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of FGP-dTV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_dTV,\ + 'input' : noisyVol,\ + 'refdata' : noisyRef,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :300 ,\ + 'tolerance_constant':0.00001,\ + 'eta_const':0.2,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("#############FGP dTV CPU####################") +start_time = timeit.default_timer() +fgp_dTV_cpu3D = FGP_dTV(pars['input'], + pars['refdata'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['eta_const'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'cpu') + + +rms = rmse(idealVol, fgp_dTV_cpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_dTV_cpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the CPU using FGP-dTV')) +#%% diff --git a/docs/demos/demo_cpu_vs_gpu_regularisers.py b/docs/demos/demo_cpu_vs_gpu_regularisers.py new file mode 100644 index 0000000..230a761 --- /dev/null +++ b/docs/demos/demo_cpu_vs_gpu_regularisers.py @@ -0,0 +1,790 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Thu Feb 22 11:39:43 2018 + +Demonstration of CPU implementation against the GPU one + +@authors: Daniil Kazantsev, Edoardo Pasca +""" + +import matplotlib.pyplot as plt +import numpy as np +import os +import timeit +from ccpi.filters.regularisers import ROF_TV, FGP_TV, SB_TV, TGV, LLT_ROF, FGP_dTV, NDF, Diff4th +from ccpi.filters.regularisers import PatchSelect +from qualitymetrics import rmse +############################################################################### +def printParametersToString(pars): + txt = r'' + for key, value in pars.items(): + if key== 'algorithm' : + txt += "{0} = {1}".format(key, value.__name__) + elif key == 'input': + txt += "{0} = {1}".format(key, np.shape(value)) + elif key == 'refdata': + txt += "{0} = {1}".format(key, np.shape(value)) + else: + txt += "{0} = {1}".format(key, value) + txt += '\n' + return txt +############################################################################### + +filename = os.path.join(".." , ".." , ".." , "data" ,"lena_gray_512.tif") + +# read image +Im = plt.imread(filename) +Im = np.asarray(Im, dtype='float32') + +Im = Im/255 +perc = 0.05 +u0 = Im + np.random.normal(loc = 0 , + scale = perc * Im , + size = np.shape(Im)) +u_ref = Im + np.random.normal(loc = 0 , + scale = 0.01 * Im , + size = np.shape(Im)) + +# map the u0 u0->u0>0 +# f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1) +u0 = u0.astype('float32') +u_ref = u_ref.astype('float32') + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________ROF-TV bench___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Comparison of ROF-TV regulariser using CPU and GPU implementations') +a=fig.add_subplot(1,4,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm': ROF_TV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04,\ + 'number_of_iterations': 4500,\ + 'time_marching_parameter': 0.00002 + } +print ("#############ROF TV CPU####################") +start_time = timeit.default_timer() +rof_cpu = ROF_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'cpu') +rms = rmse(Im, rof_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(rof_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +print ("##############ROF TV GPU##################") +start_time = timeit.default_timer() +rof_gpu = ROF_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'gpu') + +rms = rmse(Im, rof_gpu) +pars['rmse'] = rms +pars['algorithm'] = ROF_TV +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,3) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(rof_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + + +print ("--------Compare the results--------") +tolerance = 1e-05 +diff_im = np.zeros(np.shape(rof_cpu)) +diff_im = abs(rof_cpu - rof_gpu) +diff_im[diff_im > tolerance] = 1 +a=fig.add_subplot(1,4,4) +imgplot = plt.imshow(diff_im, vmin=0, vmax=1, cmap="gray") +plt.title('{}'.format('Pixels larger threshold difference')) +if (diff_im.sum() > 1): + print ("Arrays do not match!") +else: + print ("Arrays match") + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________FGP-TV bench___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Comparison of FGP-TV regulariser using CPU and GPU implementations') +a=fig.add_subplot(1,4,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_TV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :1200 ,\ + 'tolerance_constant':0.00001,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("#############FGP TV CPU####################") +start_time = timeit.default_timer() +fgp_cpu = FGP_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'cpu') + + +rms = rmse(Im, fgp_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + + +print ("##############FGP TV GPU##################") +start_time = timeit.default_timer() +fgp_gpu = FGP_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'gpu') + +rms = rmse(Im, fgp_gpu) +pars['rmse'] = rms +pars['algorithm'] = FGP_TV +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,3) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + + +print ("--------Compare the results--------") +tolerance = 1e-05 +diff_im = np.zeros(np.shape(fgp_cpu)) +diff_im = abs(fgp_cpu - fgp_gpu) +diff_im[diff_im > tolerance] = 1 +a=fig.add_subplot(1,4,4) +imgplot = plt.imshow(diff_im, vmin=0, vmax=1, cmap="gray") +plt.title('{}'.format('Pixels larger threshold difference')) +if (diff_im.sum() > 1): + print ("Arrays do not match!") +else: + print ("Arrays match") + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________SB-TV bench___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Comparison of SB-TV regulariser using CPU and GPU implementations') +a=fig.add_subplot(1,4,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : SB_TV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :150 ,\ + 'tolerance_constant':1e-05,\ + 'methodTV': 0 ,\ + 'printingOut': 0 + } + +print ("#############SB-TV CPU####################") +start_time = timeit.default_timer() +sb_cpu = SB_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['printingOut'],'cpu') + + +rms = rmse(Im, sb_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(sb_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + + +print ("##############SB TV GPU##################") +start_time = timeit.default_timer() +sb_gpu = SB_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['printingOut'],'gpu') + +rms = rmse(Im, sb_gpu) +pars['rmse'] = rms +pars['algorithm'] = SB_TV +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,3) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(sb_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +print ("--------Compare the results--------") +tolerance = 1e-05 +diff_im = np.zeros(np.shape(sb_cpu)) +diff_im = abs(sb_cpu - sb_gpu) +diff_im[diff_im > tolerance] = 1 +a=fig.add_subplot(1,4,4) +imgplot = plt.imshow(diff_im, vmin=0, vmax=1, cmap="gray") +plt.title('{}'.format('Pixels larger threshold difference')) +if (diff_im.sum() > 1): + print ("Arrays do not match!") +else: + print ("Arrays match") +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________TGV bench___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Comparison of TGV regulariser using CPU and GPU implementations') +a=fig.add_subplot(1,4,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : TGV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04, \ + 'alpha1':1.0,\ + 'alpha0':2.0,\ + 'number_of_iterations' :400 ,\ + 'LipshitzConstant' :12 ,\ + } + +print ("#############TGV CPU####################") +start_time = timeit.default_timer() +tgv_cpu = TGV(pars['input'], + pars['regularisation_parameter'], + pars['alpha1'], + pars['alpha0'], + pars['number_of_iterations'], + pars['LipshitzConstant'],'cpu') + +rms = rmse(Im, tgv_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(tgv_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +print ("##############TGV GPU##################") +start_time = timeit.default_timer() +tgv_gpu = TGV(pars['input'], + pars['regularisation_parameter'], + pars['alpha1'], + pars['alpha0'], + pars['number_of_iterations'], + pars['LipshitzConstant'],'gpu') + +rms = rmse(Im, tgv_gpu) +pars['rmse'] = rms +pars['algorithm'] = TGV +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,3) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(tgv_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +print ("--------Compare the results--------") +tolerance = 1e-05 +diff_im = np.zeros(np.shape(tgv_gpu)) +diff_im = abs(tgv_cpu - tgv_gpu) +diff_im[diff_im > tolerance] = 1 +a=fig.add_subplot(1,4,4) +imgplot = plt.imshow(diff_im, vmin=0, vmax=1, cmap="gray") +plt.title('{}'.format('Pixels larger threshold difference')) +if (diff_im.sum() > 1): + print ("Arrays do not match!") +else: + print ("Arrays match") +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________LLT-ROF bench___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Comparison of LLT-ROF regulariser using CPU and GPU implementations') +a=fig.add_subplot(1,4,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : LLT_ROF, \ + 'input' : u0,\ + 'regularisation_parameterROF':0.04, \ + 'regularisation_parameterLLT':0.01, \ + 'number_of_iterations' :4500 ,\ + 'time_marching_parameter' :0.00002 ,\ + } + +print ("#############LLT- ROF CPU####################") +start_time = timeit.default_timer() +lltrof_cpu = LLT_ROF(pars['input'], + pars['regularisation_parameterROF'], + pars['regularisation_parameterLLT'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'cpu') + +rms = rmse(Im, lltrof_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(lltrof_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +print ("#############LLT- ROF GPU####################") +start_time = timeit.default_timer() +lltrof_gpu = LLT_ROF(pars['input'], + pars['regularisation_parameterROF'], + pars['regularisation_parameterLLT'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'gpu') + +rms = rmse(Im, lltrof_gpu) +pars['rmse'] = rms +pars['algorithm'] = LLT_ROF +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,3) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(lltrof_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +print ("--------Compare the results--------") +tolerance = 1e-05 +diff_im = np.zeros(np.shape(lltrof_gpu)) +diff_im = abs(lltrof_cpu - lltrof_gpu) +diff_im[diff_im > tolerance] = 1 +a=fig.add_subplot(1,4,4) +imgplot = plt.imshow(diff_im, vmin=0, vmax=1, cmap="gray") +plt.title('{}'.format('Pixels larger threshold difference')) +if (diff_im.sum() > 1): + print ("Arrays do not match!") +else: + print ("Arrays match") +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________NDF bench___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Comparison of NDF regulariser using CPU and GPU implementations') +a=fig.add_subplot(1,4,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : NDF, \ + 'input' : u0,\ + 'regularisation_parameter':0.06, \ + 'edge_parameter':0.04,\ + 'number_of_iterations' :1000 ,\ + 'time_marching_parameter':0.025,\ + 'penalty_type': 1 + } + +print ("#############NDF CPU####################") +start_time = timeit.default_timer() +ndf_cpu = NDF(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], + pars['penalty_type'],'cpu') + +rms = rmse(Im, ndf_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(ndf_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + + +print ("##############NDF GPU##################") +start_time = timeit.default_timer() +ndf_gpu = NDF(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], + pars['penalty_type'],'gpu') + +rms = rmse(Im, ndf_gpu) +pars['rmse'] = rms +pars['algorithm'] = NDF +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,3) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(ndf_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +print ("--------Compare the results--------") +tolerance = 1e-05 +diff_im = np.zeros(np.shape(ndf_cpu)) +diff_im = abs(ndf_cpu - ndf_gpu) +diff_im[diff_im > tolerance] = 1 +a=fig.add_subplot(1,4,4) +imgplot = plt.imshow(diff_im, vmin=0, vmax=1, cmap="gray") +plt.title('{}'.format('Pixels larger threshold difference')) +if (diff_im.sum() > 1): + print ("Arrays do not match!") +else: + print ("Arrays match") + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Anisotropic Diffusion 4th Order (2D)____") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Comparison of Diff4th regulariser using CPU and GPU implementations') +a=fig.add_subplot(1,4,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : Diff4th, \ + 'input' : u0,\ + 'regularisation_parameter':3.5, \ + 'edge_parameter':0.02,\ + 'number_of_iterations' :500 ,\ + 'time_marching_parameter':0.001 + } + +print ("#############Diff4th CPU####################") +start_time = timeit.default_timer() +diff4th_cpu = Diff4th(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'cpu') + +rms = rmse(Im, diff4th_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(diff4th_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +print ("##############Diff4th GPU##################") +start_time = timeit.default_timer() +diff4th_gpu = Diff4th(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], 'gpu') + +rms = rmse(Im, diff4th_gpu) +pars['rmse'] = rms +pars['algorithm'] = Diff4th +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,3) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(diff4th_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +print ("--------Compare the results--------") +tolerance = 1e-05 +diff_im = np.zeros(np.shape(diff4th_cpu)) +diff_im = abs(diff4th_cpu - diff4th_gpu) +diff_im[diff_im > tolerance] = 1 +a=fig.add_subplot(1,4,4) +imgplot = plt.imshow(diff_im, vmin=0, vmax=1, cmap="gray") +plt.title('{}'.format('Pixels larger threshold difference')) +if (diff_im.sum() > 1): + print ("Arrays do not match!") +else: + print ("Arrays match") + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________FGP-dTV bench___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Comparison of FGP-dTV regulariser using CPU and GPU implementations') +a=fig.add_subplot(1,4,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_dTV, \ + 'input' : u0,\ + 'refdata' : u_ref,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :1000 ,\ + 'tolerance_constant':1e-07,\ + 'eta_const':0.2,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("#############FGP dTV CPU####################") +start_time = timeit.default_timer() +fgp_dtv_cpu = FGP_dTV(pars['input'], + pars['refdata'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['eta_const'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'cpu') + + +rms = rmse(Im, fgp_dtv_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_dtv_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) + +print ("##############FGP dTV GPU##################") +start_time = timeit.default_timer() +fgp_dtv_gpu = FGP_dTV(pars['input'], + pars['refdata'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['eta_const'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'gpu') +rms = rmse(Im, fgp_dtv_gpu) +pars['rmse'] = rms +pars['algorithm'] = FGP_dTV +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,4,3) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_dtv_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + + +print ("--------Compare the results--------") +tolerance = 1e-05 +diff_im = np.zeros(np.shape(fgp_dtv_cpu)) +diff_im = abs(fgp_dtv_cpu - fgp_dtv_gpu) +diff_im[diff_im > tolerance] = 1 +a=fig.add_subplot(1,4,4) +imgplot = plt.imshow(diff_im, vmin=0, vmax=1, cmap="gray") +plt.title('{}'.format('Pixels larger threshold difference')) +if (diff_im.sum() > 1): + print ("Arrays do not match!") +else: + print ("Arrays match") +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____Non-local regularisation bench_________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Comparison of Nonlocal TV regulariser using CPU and GPU implementations') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +pars = {'algorithm' : PatchSelect, \ + 'input' : u0,\ + 'searchwindow': 7, \ + 'patchwindow': 2,\ + 'neighbours' : 15 ,\ + 'edge_parameter':0.18} + +print ("############## Nonlocal Patches on CPU##################") +start_time = timeit.default_timer() +H_i, H_j, WeightsCPU = PatchSelect(pars['input'], + pars['searchwindow'], + pars['patchwindow'], + pars['neighbours'], + pars['edge_parameter'],'cpu') +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) + +print ("############## Nonlocal Patches on GPU##################") +start_time = timeit.default_timer() +start_time = timeit.default_timer() +H_i, H_j, WeightsGPU = PatchSelect(pars['input'], + pars['searchwindow'], + pars['patchwindow'], + pars['neighbours'], + pars['edge_parameter'],'gpu') +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) + +print ("--------Compare the results--------") +tolerance = 1e-05 +diff_im = np.zeros(np.shape(u0)) +diff_im = abs(WeightsCPU[0,:,:] - WeightsGPU[0,:,:]) +diff_im[diff_im > tolerance] = 1 +a=fig.add_subplot(1,2,2) +imgplot = plt.imshow(diff_im, vmin=0, vmax=1, cmap="gray") +plt.title('{}'.format('Pixels larger threshold difference')) +if (diff_im.sum() > 1): + print ("Arrays do not match!") +else: + print ("Arrays match") +#%%
\ No newline at end of file diff --git a/docs/demos/demo_gpu_regularisers.py b/docs/demos/demo_gpu_regularisers.py new file mode 100644 index 0000000..e1c6575 --- /dev/null +++ b/docs/demos/demo_gpu_regularisers.py @@ -0,0 +1,518 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Thu Feb 22 11:39:43 2018 + +Demonstration of GPU regularisers + +@authors: Daniil Kazantsev, Edoardo Pasca +""" + +import matplotlib.pyplot as plt +import numpy as np +import os +import timeit +from ccpi.filters.regularisers import ROF_TV, FGP_TV, SB_TV, TGV, LLT_ROF, FGP_dTV, NDF, Diff4th +from ccpi.filters.regularisers import PatchSelect, NLTV +from qualitymetrics import rmse +############################################################################### +def printParametersToString(pars): + txt = r'' + for key, value in pars.items(): + if key== 'algorithm' : + txt += "{0} = {1}".format(key, value.__name__) + elif key == 'input': + txt += "{0} = {1}".format(key, np.shape(value)) + elif key == 'refdata': + txt += "{0} = {1}".format(key, np.shape(value)) + else: + txt += "{0} = {1}".format(key, value) + txt += '\n' + return txt +############################################################################### +#%% +filename = os.path.join(".." , ".." , ".." , "data" ,"lena_gray_512.tif") + +# read image +Im = plt.imread(filename) +Im = np.asarray(Im, dtype='float32') + +Im = Im/255 +perc = 0.05 +u0 = Im + np.random.normal(loc = 0 , + scale = perc * Im , + size = np.shape(Im)) +u_ref = Im + np.random.normal(loc = 0 , + scale = 0.01 * Im , + size = np.shape(Im)) +(N,M) = np.shape(u0) +# map the u0 u0->u0>0 +# f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1) +u0 = u0.astype('float32') +u_ref = u_ref.astype('float32') +""" +M = M-100 +u_ref2 = np.zeros([N,M],dtype='float32') +u_ref2[:,0:M] = u_ref[:,0:M] +u_ref = u_ref2 +del u_ref2 + +u02 = np.zeros([N,M],dtype='float32') +u02[:,0:M] = u0[:,0:M] +u0 = u02 +del u02 + +Im2 = np.zeros([N,M],dtype='float32') +Im2[:,0:M] = Im[:,0:M] +Im = Im2 +del Im2 +""" +#%% + +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________ROF-TV regulariser_____________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of the ROF-TV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm': ROF_TV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04,\ + 'number_of_iterations': 1200,\ + 'time_marching_parameter': 0.0025 + } +print ("##############ROF TV GPU##################") +start_time = timeit.default_timer() +rof_gpu = ROF_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'gpu') + +rms = rmse(Im, rof_gpu) +pars['rmse'] = rms +pars['algorithm'] = ROF_TV +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(rof_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________FGP-TV regulariser_____________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of the FGP-TV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_TV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :1200 ,\ + 'tolerance_constant':1e-06,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("##############FGP TV GPU##################") +start_time = timeit.default_timer() +fgp_gpu = FGP_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'gpu') + +rms = rmse(Im, fgp_gpu) +pars['rmse'] = rms +pars['algorithm'] = FGP_TV +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________SB-TV regulariser______________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of the SB-TV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : SB_TV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :150 ,\ + 'tolerance_constant':1e-06,\ + 'methodTV': 0 ,\ + 'printingOut': 0 + } + +print ("##############SB TV GPU##################") +start_time = timeit.default_timer() +sb_gpu = SB_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['printingOut'],'gpu') + +rms = rmse(Im, sb_gpu) +pars['rmse'] = rms +pars['algorithm'] = SB_TV +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(sb_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) +#%% + +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_____Total Generalised Variation (2D)______") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of TGV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : TGV, \ + 'input' : u0,\ + 'regularisation_parameter':0.04, \ + 'alpha1':1.0,\ + 'alpha0':2.0,\ + 'number_of_iterations' :1250 ,\ + 'LipshitzConstant' :12 ,\ + } + +print ("#############TGV CPU####################") +start_time = timeit.default_timer() +tgv_gpu = TGV(pars['input'], + pars['regularisation_parameter'], + pars['alpha1'], + pars['alpha0'], + pars['number_of_iterations'], + pars['LipshitzConstant'],'gpu') + + +rms = rmse(Im, tgv_gpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(tgv_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +#%% + +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("______________LLT- ROF (2D)________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of LLT-ROF regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : LLT_ROF, \ + 'input' : u0,\ + 'regularisation_parameterROF':0.04, \ + 'regularisation_parameterLLT':0.01, \ + 'number_of_iterations' :500 ,\ + 'time_marching_parameter' :0.0025 ,\ + } + +print ("#############LLT- ROF GPU####################") +start_time = timeit.default_timer() +lltrof_gpu = LLT_ROF(pars['input'], + pars['regularisation_parameterROF'], + pars['regularisation_parameterLLT'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'gpu') + + +rms = rmse(Im, lltrof_gpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(lltrof_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________NDF regulariser_____________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of the NDF regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : NDF, \ + 'input' : u0,\ + 'regularisation_parameter':0.025, \ + 'edge_parameter':0.015,\ + 'number_of_iterations' :500 ,\ + 'time_marching_parameter':0.025,\ + 'penalty_type': 1 + } + +print ("##############NDF GPU##################") +start_time = timeit.default_timer() +ndf_gpu = NDF(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], + pars['penalty_type'],'gpu') + +rms = rmse(Im, ndf_gpu) +pars['rmse'] = rms +pars['algorithm'] = NDF +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(ndf_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Anisotropic Diffusion 4th Order (2D)____") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of Diff4th regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : Diff4th, \ + 'input' : u0,\ + 'regularisation_parameter':3.5, \ + 'edge_parameter':0.02,\ + 'number_of_iterations' :500 ,\ + 'time_marching_parameter':0.0015 + } + +print ("#############DIFF4th CPU################") +start_time = timeit.default_timer() +diff4_gpu = Diff4th(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'gpu') + +rms = rmse(Im, diff4_gpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(diff4_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Nonlocal patches pre-calculation____") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +start_time = timeit.default_timer() +# set parameters +pars = {'algorithm' : PatchSelect, \ + 'input' : u0,\ + 'searchwindow': 7, \ + 'patchwindow': 2,\ + 'neighbours' : 15 ,\ + 'edge_parameter':0.18} + +H_i, H_j, Weights = PatchSelect(pars['input'], + pars['searchwindow'], + pars['patchwindow'], + pars['neighbours'], + pars['edge_parameter'],'gpu') + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +""" +plt.figure() +plt.imshow(Weights[0,:,:],cmap="gray",interpolation="nearest",vmin=0, vmax=1) +plt.show() +""" +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Nonlocal Total Variation penalty____") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +## plot +fig = plt.figure() +plt.suptitle('Performance of NLTV regulariser using the CPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +pars2 = {'algorithm' : NLTV, \ + 'input' : u0,\ + 'H_i': H_i, \ + 'H_j': H_j,\ + 'H_k' : 0,\ + 'Weights' : Weights,\ + 'regularisation_parameter': 0.02,\ + 'iterations': 3 + } +start_time = timeit.default_timer() +nltv_cpu = NLTV(pars2['input'], + pars2['H_i'], + pars2['H_j'], + pars2['H_k'], + pars2['Weights'], + pars2['regularisation_parameter'], + pars2['iterations']) + +rms = rmse(Im, nltv_cpu) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(nltv_cpu, cmap="gray") +plt.title('{}'.format('CPU results')) +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("____________FGP-dTV bench___________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of the FGP-dTV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(u0,cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_dTV, \ + 'input' : u0,\ + 'refdata' : u_ref,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :2000 ,\ + 'tolerance_constant':1e-06,\ + 'eta_const':0.2,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("##############FGP dTV GPU##################") +start_time = timeit.default_timer() +fgp_dtv_gpu = FGP_dTV(pars['input'], + pars['refdata'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['eta_const'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'gpu') + +rms = rmse(Im, fgp_dtv_gpu) +pars['rmse'] = rms +pars['algorithm'] = FGP_dTV +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_dtv_gpu, cmap="gray") +plt.title('{}'.format('GPU results')) diff --git a/docs/demos/demo_gpu_regularisers3D.py b/docs/demos/demo_gpu_regularisers3D.py new file mode 100644 index 0000000..b6058d2 --- /dev/null +++ b/docs/demos/demo_gpu_regularisers3D.py @@ -0,0 +1,460 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Thu Feb 22 11:39:43 2018 + +Demonstration of GPU regularisers + +@authors: Daniil Kazantsev, Edoardo Pasca +""" + +import matplotlib.pyplot as plt +import numpy as np +import os +import timeit +from ccpi.filters.regularisers import ROF_TV, FGP_TV, SB_TV, TGV, LLT_ROF, FGP_dTV, NDF, Diff4th +from qualitymetrics import rmse +############################################################################### +def printParametersToString(pars): + txt = r'' + for key, value in pars.items(): + if key== 'algorithm' : + txt += "{0} = {1}".format(key, value.__name__) + elif key == 'input': + txt += "{0} = {1}".format(key, np.shape(value)) + elif key == 'refdata': + txt += "{0} = {1}".format(key, np.shape(value)) + else: + txt += "{0} = {1}".format(key, value) + txt += '\n' + return txt +############################################################################### +#%% +filename = os.path.join(".." , ".." , ".." , "data" ,"lena_gray_512.tif") + +# read image +Im = plt.imread(filename) +Im = np.asarray(Im, dtype='float32') + +Im = Im/255 +perc = 0.05 +u0 = Im + np.random.normal(loc = 0 , + scale = perc * Im , + size = np.shape(Im)) +u_ref = Im + np.random.normal(loc = 0 , + scale = 0.01 * Im , + size = np.shape(Im)) +(N,M) = np.shape(u0) +# map the u0 u0->u0>0 +# f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1) +u0 = u0.astype('float32') +u_ref = u_ref.astype('float32') +""" +M = M-100 +u_ref2 = np.zeros([N,M],dtype='float32') +u_ref2[:,0:M] = u_ref[:,0:M] +u_ref = u_ref2 +del u_ref2 + +u02 = np.zeros([N,M],dtype='float32') +u02[:,0:M] = u0[:,0:M] +u0 = u02 +del u02 + +Im2 = np.zeros([N,M],dtype='float32') +Im2[:,0:M] = Im[:,0:M] +Im = Im2 +del Im2 +""" + + +slices = 20 + +filename = os.path.join(".." , ".." , ".." , "data" ,"lena_gray_512.tif") +Im = plt.imread(filename) +Im = np.asarray(Im, dtype='float32') + +Im = Im/255 +perc = 0.05 + +noisyVol = np.zeros((slices,N,N),dtype='float32') +noisyRef = np.zeros((slices,N,N),dtype='float32') +idealVol = np.zeros((slices,N,N),dtype='float32') + +for i in range (slices): + noisyVol[i,:,:] = Im + np.random.normal(loc = 0 , scale = perc * Im , size = np.shape(Im)) + noisyRef[i,:,:] = Im + np.random.normal(loc = 0 , scale = 0.01 * Im , size = np.shape(Im)) + idealVol[i,:,:] = Im + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________ROF-TV (3D)_________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of ROF-TV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy 15th slice of a volume') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm': ROF_TV, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.04,\ + 'number_of_iterations': 500,\ + 'time_marching_parameter': 0.0025 + } +print ("#############ROF TV GPU####################") +start_time = timeit.default_timer() +rof_gpu3D = ROF_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'gpu') +rms = rmse(idealVol, rof_gpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(rof_gpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the GPU using ROF-TV')) +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________FGP-TV (3D)__________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of FGP-TV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_TV, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :300 ,\ + 'tolerance_constant':0.00001,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("#############FGP TV GPU####################") +start_time = timeit.default_timer() +fgp_gpu3D = FGP_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'gpu') + +rms = rmse(idealVol, fgp_gpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_gpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the GPU using FGP-TV')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________SB-TV (3D)__________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of SB-TV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : SB_TV, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :100 ,\ + 'tolerance_constant':1e-05,\ + 'methodTV': 0 ,\ + 'printingOut': 0 + } + +print ("#############SB TV GPU####################") +start_time = timeit.default_timer() +sb_gpu3D = SB_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['methodTV'], + pars['printingOut'],'gpu') + +rms = rmse(idealVol, sb_gpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(sb_gpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the GPU using SB-TV')) +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________LLT-ROF (3D)_________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of LLT-ROF regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : LLT_ROF, \ + 'input' : noisyVol,\ + 'regularisation_parameterROF':0.04, \ + 'regularisation_parameterLLT':0.015, \ + 'number_of_iterations' :300 ,\ + 'time_marching_parameter' :0.0025 ,\ + } + +print ("#############LLT ROF CPU####################") +start_time = timeit.default_timer() +lltrof_gpu3D = LLT_ROF(pars['input'], + pars['regularisation_parameterROF'], + pars['regularisation_parameterLLT'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'gpu') + +rms = rmse(idealVol, lltrof_gpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(lltrof_gpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the GPU using LLT-ROF')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________TGV (3D)_________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of TGV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : TGV, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.04, \ + 'alpha1':1.0,\ + 'alpha0':2.0,\ + 'number_of_iterations' :600 ,\ + 'LipshitzConstant' :12 ,\ + } + +print ("#############TGV GPU####################") +start_time = timeit.default_timer() +tgv_gpu3D = TGV(pars['input'], + pars['regularisation_parameter'], + pars['alpha1'], + pars['alpha0'], + pars['number_of_iterations'], + pars['LipshitzConstant'],'gpu') + + +rms = rmse(idealVol, tgv_gpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(tgv_gpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the GPU using TGV')) +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________NDF-TV (3D)_________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of NDF regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : NDF, \ + 'input' : noisyVol,\ + 'regularisation_parameter':0.025, \ + 'edge_parameter':0.015,\ + 'number_of_iterations' :500 ,\ + 'time_marching_parameter':0.025,\ + 'penalty_type': 1 + } + +print ("#############NDF GPU####################") +start_time = timeit.default_timer() +ndf_gpu3D = NDF(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], + pars['penalty_type'],'gpu') + +rms = rmse(idealVol, ndf_gpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(ndf_gpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the GPU using NDF')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("___Anisotropic Diffusion 4th Order (3D)____") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of DIFF4th regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : Diff4th, \ + 'input' : noisyVol,\ + 'regularisation_parameter':3.5, \ + 'edge_parameter':0.02,\ + 'number_of_iterations' :300 ,\ + 'time_marching_parameter':0.0015 + } + +print ("#############DIFF4th CPU################") +start_time = timeit.default_timer() +diff4_gpu3D = Diff4th(pars['input'], + pars['regularisation_parameter'], + pars['edge_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'],'gpu') + +rms = rmse(idealVol, diff4_gpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(diff4_gpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('GPU results')) + +#%% +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") +print ("_______________FGP-dTV (3D)________________") +print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + +## plot +fig = plt.figure() +plt.suptitle('Performance of FGP-dTV regulariser using the GPU') +a=fig.add_subplot(1,2,1) +a.set_title('Noisy Image') +imgplot = plt.imshow(noisyVol[10,:,:],cmap="gray") + +# set parameters +pars = {'algorithm' : FGP_dTV, \ + 'input' : noisyVol,\ + 'refdata' : noisyRef,\ + 'regularisation_parameter':0.04, \ + 'number_of_iterations' :300 ,\ + 'tolerance_constant':0.00001,\ + 'eta_const':0.2,\ + 'methodTV': 0 ,\ + 'nonneg': 0 ,\ + 'printingOut': 0 + } + +print ("#############FGP TV GPU####################") +start_time = timeit.default_timer() +fgp_dTV_gpu3D = FGP_dTV(pars['input'], + pars['refdata'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['tolerance_constant'], + pars['eta_const'], + pars['methodTV'], + pars['nonneg'], + pars['printingOut'],'gpu') + +rms = rmse(idealVol, fgp_dTV_gpu3D) +pars['rmse'] = rms + +txtstr = printParametersToString(pars) +txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time) +print (txtstr) +a=fig.add_subplot(1,2,2) + +# these are matplotlib.patch.Patch properties +props = dict(boxstyle='round', facecolor='wheat', alpha=0.75) +# place a text box in upper left in axes coords +a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14, + verticalalignment='top', bbox=props) +imgplot = plt.imshow(fgp_dTV_gpu3D[10,:,:], cmap="gray") +plt.title('{}'.format('Recovered volume on the GPU using FGP-dTV')) +#%% diff --git a/docs/demos/qualitymetrics.py b/docs/demos/qualitymetrics.py new file mode 100644 index 0000000..850829e --- /dev/null +++ b/docs/demos/qualitymetrics.py @@ -0,0 +1,18 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +""" +Created on Wed Feb 21 13:34:32 2018 +# quality metrics +@authors: Daniil Kazantsev, Edoardo Pasca +""" +import numpy as np + +def nrmse(im1, im2): + rmse = np.sqrt(np.sum((im2 - im1) ** 2) / float(im1.size)) + max_val = max(np.max(im1), np.max(im2)) + min_val = min(np.min(im1), np.min(im2)) + return 1 - (rmse / (max_val - min_val)) + +def rmse(im1, im2): + rmse = np.sqrt(np.sum((im1 - im2) ** 2) / float(im1.size)) + return rmse diff --git a/recipe/bld.bat b/recipe/bld.bat new file mode 100644 index 0000000..6c84355 --- /dev/null +++ b/recipe/bld.bat @@ -0,0 +1,20 @@ +IF NOT DEFINED CIL_VERSION ( +ECHO CIL_VERSION Not Defined. +exit 1 +) + +mkdir "%SRC_DIR%\ccpi" +ROBOCOPY /E "%RECIPE_DIR%\..\.." "%SRC_DIR%\ccpi" +ROBOCOPY /E "%RECIPE_DIR%\..\..\..\Core" "%SRC_DIR%\Core" +::cd %SRC_DIR%\ccpi\Python +cd %SRC_DIR% + +:: issue cmake to create setup.py +cmake -G "NMake Makefiles" %RECIPE_DIR%\..\..\..\ -DBUILD_PYTHON_WRAPPERS=ON -DCONDA_BUILD=ON -DBUILD_CUDA=OFF -DCMAKE_BUILD_TYPE="Release" -DLIBRARY_LIB="%CONDA_PREFIX%\lib" -DLIBRARY_INC="%CONDA_PREFIX%" -DCMAKE_INSTALL_PREFIX="%PREFIX%\Library" + +::%PYTHON% setup-regularisers.py build_ext +::if errorlevel 1 exit 1 +::%PYTHON% setup-regularisers.py install +::if errorlevel 1 exit 1 +nmake install +if errorlevel 1 exit 1
\ No newline at end of file diff --git a/recipe/build.sh b/recipe/build.sh new file mode 100644 index 0000000..1d54b6f --- /dev/null +++ b/recipe/build.sh @@ -0,0 +1,18 @@ + +mkdir "$SRC_DIR/ccpi" +cp -rv "$RECIPE_DIR/../src/Matlab" "$SRC_DIR/ccpi" +cp -rv "$RECIPE_DIR/../src/Python" "$SRC_DIR/ccpi" +cp -rv "$RECIPE_DIR/../src/Core" "$SRC_DIR/Core" + +cd $SRC_DIR +##cuda=off + +cmake -G "Unix Makefiles" $RECIPE_DIR/../ -DBUILD_PYTHON_WRAPPER=ON -DCONDA_BUILD=ON -DBUILD_CUDA=ON -DCMAKE_BUILD_TYPE="Release" -DLIBRARY_LIB=$CONDA_PREFIX/lib -DLIBRARY_INC=$CONDA_PREFIX -DCMAKE_INSTALL_PREFIX=$PREFIX + + +make install + +#$PYTHON setup-regularisers.py build_ext +#$PYTHON setup-regularisers.py install + + diff --git a/recipe/conda_build_config.yaml b/recipe/conda_build_config.yaml new file mode 100644 index 0000000..fbe82dc --- /dev/null +++ b/recipe/conda_build_config.yaml @@ -0,0 +1,9 @@ +python: + - 2.7 # [not win] + - 3.5 + - 3.6 +# - 3.7 +numpy: + - 1.12 + - 1.14 + - 1.15 diff --git a/recipe/meta.yaml b/recipe/meta.yaml new file mode 100644 index 0000000..7435b2b --- /dev/null +++ b/recipe/meta.yaml @@ -0,0 +1,40 @@ +package: + name: ccpi-regulariser + version: {{CIL_VERSION}} + +build: + preserve_egg_dir: False + number: 0 + script_env: + - CIL_VERSION + +test: + files: + - lena_gray_512.tif + requires: + - pillow=4.1.1 + +requirements: + build: + - python + - numpy {{ numpy }} + - setuptools + - cython + - vc 14 # [win and py36] + - vc 14 # [win and py35] + - vc 9 # [win and py27] + - cmake + + run: + - {{ pin_compatible('numpy', max_pin='x.x') }} + - python + - numpy + - vc 14 # [win and py36] + - vc 14 # [win and py35] + - vc 9 # [win and py27] + - libgcc-ng + +about: + home: http://www.ccpi.ac.uk + license: BSD license + summary: 'CCPi Core Imaging Library Quantification Toolbox' diff --git a/recipe/run_test.py b/recipe/run_test.py new file mode 100755 index 0000000..21f3216 --- /dev/null +++ b/recipe/run_test.py @@ -0,0 +1,819 @@ +import unittest
+import numpy as np
+import os
+import timeit
+from ccpi.filters.regularisers import ROF_TV, FGP_TV, SB_TV, TGV, LLT_ROF, FGP_dTV, NDF, Diff4th
+from PIL import Image
+
+class TiffReader(object):
+ def imread(self, filename):
+ return np.asarray(Image.open(filename))
+###############################################################################
+def printParametersToString(pars):
+ txt = r''
+ for key, value in pars.items():
+ if key== 'algorithm' :
+ txt += "{0} = {1}".format(key, value.__name__)
+ elif key == 'input':
+ txt += "{0} = {1}".format(key, np.shape(value))
+ elif key == 'refdata':
+ txt += "{0} = {1}".format(key, np.shape(value))
+ else:
+ txt += "{0} = {1}".format(key, value)
+ txt += '\n'
+ return txt
+def nrmse(im1, im2):
+ rmse = np.sqrt(np.sum((im2 - im1) ** 2) / float(im1.size))
+ max_val = max(np.max(im1), np.max(im2))
+ min_val = min(np.min(im1), np.min(im2))
+ return 1 - (rmse / (max_val - min_val))
+
+def rmse(im1, im2):
+ rmse = np.sqrt(np.sum((im1 - im2) ** 2) / float(im1.size))
+ return rmse
+###############################################################################
+
+class TestRegularisers(unittest.TestCase):
+
+
+ def test_ROF_TV_CPU_vs_GPU(self):
+ #print ("tomas debug test function")
+ print(__name__)
+ filename = os.path.join("lena_gray_512.tif")
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+
+ Im = Im/255
+ perc = 0.05
+ u0 = Im + np.random.normal(loc = 0 ,
+ scale = perc * Im ,
+ size = np.shape(Im))
+ u_ref = Im + np.random.normal(loc = 0 ,
+ scale = 0.01 * Im ,
+ size = np.shape(Im))
+
+ # map the u0 u0->u0>0
+ # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1)
+ u0 = u0.astype('float32')
+ u_ref = u_ref.astype('float32')
+
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("____________ROF-TV bench___________________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+
+ # set parameters
+ pars = {'algorithm': ROF_TV, \
+ 'input' : u0,\
+ 'regularisation_parameter':0.04,\
+ 'number_of_iterations': 2500,\
+ 'time_marching_parameter': 0.00002
+ }
+ print ("#############ROF TV CPU####################")
+ start_time = timeit.default_timer()
+ rof_cpu = ROF_TV(pars['input'],
+ pars['regularisation_parameter'],
+ pars['number_of_iterations'],
+ pars['time_marching_parameter'],'cpu')
+ rms = rmse(Im, rof_cpu)
+ pars['rmse'] = rms
+
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("##############ROF TV GPU##################")
+ start_time = timeit.default_timer()
+ try:
+ rof_gpu = ROF_TV(pars['input'],
+ pars['regularisation_parameter'],
+ pars['number_of_iterations'],
+ pars['time_marching_parameter'],'gpu')
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+
+ rms = rmse(Im, rof_gpu)
+ pars['rmse'] = rms
+ pars['algorithm'] = ROF_TV
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("--------Compare the results--------")
+ tolerance = 1e-04
+ diff_im = np.zeros(np.shape(rof_cpu))
+ diff_im = abs(rof_cpu - rof_gpu)
+ diff_im[diff_im > tolerance] = 1
+ self.assertLessEqual(diff_im.sum() , 1)
+
+ def test_FGP_TV_CPU_vs_GPU(self):
+ print(__name__)
+ filename = os.path.join("lena_gray_512.tif")
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+
+ Im = Im/255
+ perc = 0.05
+ u0 = Im + np.random.normal(loc = 0 ,
+ scale = perc * Im ,
+ size = np.shape(Im))
+ u_ref = Im + np.random.normal(loc = 0 ,
+ scale = 0.01 * Im ,
+ size = np.shape(Im))
+
+ # map the u0 u0->u0>0
+ # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1)
+ u0 = u0.astype('float32')
+ u_ref = u_ref.astype('float32')
+
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("____________FGP-TV bench___________________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+
+
+ # set parameters
+ pars = {'algorithm' : FGP_TV, \
+ 'input' : u0,\
+ 'regularisation_parameter':0.04, \
+ 'number_of_iterations' :1200 ,\
+ 'tolerance_constant':0.00001,\
+ 'methodTV': 0 ,\
+ 'nonneg': 0 ,\
+ 'printingOut': 0
+ }
+
+ print ("#############FGP TV CPU####################")
+ start_time = timeit.default_timer()
+ fgp_cpu = FGP_TV(pars['input'],
+ pars['regularisation_parameter'],
+ pars['number_of_iterations'],
+ pars['tolerance_constant'],
+ pars['methodTV'],
+ pars['nonneg'],
+ pars['printingOut'],'cpu')
+
+
+ rms = rmse(Im, fgp_cpu)
+ pars['rmse'] = rms
+
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+
+ print ("##############FGP TV GPU##################")
+ start_time = timeit.default_timer()
+ try:
+ fgp_gpu = FGP_TV(pars['input'],
+ pars['regularisation_parameter'],
+ pars['number_of_iterations'],
+ pars['tolerance_constant'],
+ pars['methodTV'],
+ pars['nonneg'],
+ pars['printingOut'],'gpu')
+
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+
+ rms = rmse(Im, fgp_gpu)
+ pars['rmse'] = rms
+ pars['algorithm'] = FGP_TV
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+
+ print ("--------Compare the results--------")
+ tolerance = 1e-05
+ diff_im = np.zeros(np.shape(fgp_cpu))
+ diff_im = abs(fgp_cpu - fgp_gpu)
+ diff_im[diff_im > tolerance] = 1
+
+ self.assertLessEqual(diff_im.sum() , 1)
+
+ def test_SB_TV_CPU_vs_GPU(self):
+ print(__name__)
+ filename = os.path.join("lena_gray_512.tif")
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+
+ Im = Im/255
+ perc = 0.05
+ u0 = Im + np.random.normal(loc = 0 ,
+ scale = perc * Im ,
+ size = np.shape(Im))
+ u_ref = Im + np.random.normal(loc = 0 ,
+ scale = 0.01 * Im ,
+ size = np.shape(Im))
+
+ # map the u0 u0->u0>0
+ # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1)
+ u0 = u0.astype('float32')
+ u_ref = u_ref.astype('float32')
+
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("____________SB-TV bench___________________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+
+
+ # set parameters
+ pars = {'algorithm' : SB_TV, \
+ 'input' : u0,\
+ 'regularisation_parameter':0.04, \
+ 'number_of_iterations' :150 ,\
+ 'tolerance_constant':1e-05,\
+ 'methodTV': 0 ,\
+ 'printingOut': 0
+ }
+
+ print ("#############SB-TV CPU####################")
+ start_time = timeit.default_timer()
+ sb_cpu = SB_TV(pars['input'],
+ pars['regularisation_parameter'],
+ pars['number_of_iterations'],
+ pars['tolerance_constant'],
+ pars['methodTV'],
+ pars['printingOut'],'cpu')
+
+
+ rms = rmse(Im, sb_cpu)
+ pars['rmse'] = rms
+
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+
+ print ("##############SB TV GPU##################")
+ start_time = timeit.default_timer()
+ try:
+
+ sb_gpu = SB_TV(pars['input'],
+ pars['regularisation_parameter'],
+ pars['number_of_iterations'],
+ pars['tolerance_constant'],
+ pars['methodTV'],
+ pars['printingOut'],'gpu')
+
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+
+ rms = rmse(Im, sb_gpu)
+ pars['rmse'] = rms
+ pars['algorithm'] = SB_TV
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("--------Compare the results--------")
+ tolerance = 1e-05
+ diff_im = np.zeros(np.shape(sb_cpu))
+ diff_im = abs(sb_cpu - sb_gpu)
+ diff_im[diff_im > tolerance] = 1
+ self.assertLessEqual(diff_im.sum(), 1)
+
+ def test_TGV_CPU_vs_GPU(self):
+ print(__name__)
+ filename = os.path.join("lena_gray_512.tif")
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+
+ Im = Im/255
+ perc = 0.05
+ u0 = Im + np.random.normal(loc = 0 ,
+ scale = perc * Im ,
+ size = np.shape(Im))
+ u_ref = Im + np.random.normal(loc = 0 ,
+ scale = 0.01 * Im ,
+ size = np.shape(Im))
+
+ # map the u0 u0->u0>0
+ # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1)
+ u0 = u0.astype('float32')
+ u_ref = u_ref.astype('float32')
+
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("____________TGV bench___________________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+
+
+ # set parameters
+ pars = {'algorithm' : TGV, \
+ 'input' : u0,\
+ 'regularisation_parameter':0.04, \
+ 'alpha1':1.0,\
+ 'alpha0':2.0,\
+ 'number_of_iterations' :250 ,\
+ 'LipshitzConstant' :12 ,\
+ }
+
+ print ("#############TGV CPU####################")
+ start_time = timeit.default_timer()
+ tgv_cpu = TGV(pars['input'],
+ pars['regularisation_parameter'],
+ pars['alpha1'],
+ pars['alpha0'],
+ pars['number_of_iterations'],
+ pars['LipshitzConstant'],'cpu')
+
+ rms = rmse(Im, tgv_cpu)
+ pars['rmse'] = rms
+
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+
+ print ("##############TGV GPU##################")
+ start_time = timeit.default_timer()
+ try:
+ tgv_gpu = TGV(pars['input'],
+ pars['regularisation_parameter'],
+ pars['alpha1'],
+ pars['alpha0'],
+ pars['number_of_iterations'],
+ pars['LipshitzConstant'],'gpu')
+
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+
+ rms = rmse(Im, tgv_gpu)
+ pars['rmse'] = rms
+ pars['algorithm'] = TGV
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("--------Compare the results--------")
+ tolerance = 1e-05
+ diff_im = np.zeros(np.shape(tgv_gpu))
+ diff_im = abs(tgv_cpu - tgv_gpu)
+ diff_im[diff_im > tolerance] = 1
+ self.assertLessEqual(diff_im.sum() , 1)
+
+ def test_LLT_ROF_CPU_vs_GPU(self):
+ print(__name__)
+ filename = os.path.join("lena_gray_512.tif")
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+
+ Im = Im/255
+ perc = 0.05
+ u0 = Im + np.random.normal(loc = 0 ,
+ scale = perc * Im ,
+ size = np.shape(Im))
+ u_ref = Im + np.random.normal(loc = 0 ,
+ scale = 0.01 * Im ,
+ size = np.shape(Im))
+
+ # map the u0 u0->u0>0
+ # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1)
+ u0 = u0.astype('float32')
+ u_ref = u_ref.astype('float32')
+
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("____________LLT-ROF bench___________________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+
+
+ # set parameters
+ pars = {'algorithm' : LLT_ROF, \
+ 'input' : u0,\
+ 'regularisation_parameterROF':0.04, \
+ 'regularisation_parameterLLT':0.01, \
+ 'number_of_iterations' :1000 ,\
+ 'time_marching_parameter' :0.0001 ,\
+ }
+
+ print ("#############LLT- ROF CPU####################")
+ start_time = timeit.default_timer()
+ lltrof_cpu = LLT_ROF(pars['input'],
+ pars['regularisation_parameterROF'],
+ pars['regularisation_parameterLLT'],
+ pars['number_of_iterations'],
+ pars['time_marching_parameter'],'cpu')
+
+ rms = rmse(Im, lltrof_cpu)
+ pars['rmse'] = rms
+
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("#############LLT- ROF GPU####################")
+ start_time = timeit.default_timer()
+ try:
+ lltrof_gpu = LLT_ROF(pars['input'],
+ pars['regularisation_parameterROF'],
+ pars['regularisation_parameterLLT'],
+ pars['number_of_iterations'],
+ pars['time_marching_parameter'],'gpu')
+
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+
+ rms = rmse(Im, lltrof_gpu)
+ pars['rmse'] = rms
+ pars['algorithm'] = LLT_ROF
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("--------Compare the results--------")
+ tolerance = 1e-04
+ diff_im = np.zeros(np.shape(lltrof_gpu))
+ diff_im = abs(lltrof_cpu - lltrof_gpu)
+ diff_im[diff_im > tolerance] = 1
+ self.assertLessEqual(diff_im.sum(), 1)
+
+ def test_NDF_CPU_vs_GPU(self):
+ print(__name__)
+ filename = os.path.join("lena_gray_512.tif")
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+
+ Im = Im/255
+ perc = 0.05
+ u0 = Im + np.random.normal(loc = 0 ,
+ scale = perc * Im ,
+ size = np.shape(Im))
+ u_ref = Im + np.random.normal(loc = 0 ,
+ scale = 0.01 * Im ,
+ size = np.shape(Im))
+
+ # map the u0 u0->u0>0
+ # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1)
+ u0 = u0.astype('float32')
+ u_ref = u_ref.astype('float32')
+
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("_______________NDF bench___________________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+
+
+ # set parameters
+ pars = {'algorithm' : NDF, \
+ 'input' : u0,\
+ 'regularisation_parameter':0.06, \
+ 'edge_parameter':0.04,\
+ 'number_of_iterations' :1000 ,\
+ 'time_marching_parameter':0.025,\
+ 'penalty_type': 1
+ }
+
+ print ("#############NDF CPU####################")
+ start_time = timeit.default_timer()
+ ndf_cpu = NDF(pars['input'],
+ pars['regularisation_parameter'],
+ pars['edge_parameter'],
+ pars['number_of_iterations'],
+ pars['time_marching_parameter'],
+ pars['penalty_type'],'cpu')
+
+ rms = rmse(Im, ndf_cpu)
+ pars['rmse'] = rms
+
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+
+ print ("##############NDF GPU##################")
+ start_time = timeit.default_timer()
+ try:
+ ndf_gpu = NDF(pars['input'],
+ pars['regularisation_parameter'],
+ pars['edge_parameter'],
+ pars['number_of_iterations'],
+ pars['time_marching_parameter'],
+ pars['penalty_type'],'gpu')
+
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+ rms = rmse(Im, ndf_gpu)
+ pars['rmse'] = rms
+ pars['algorithm'] = NDF
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("--------Compare the results--------")
+ tolerance = 1e-05
+ diff_im = np.zeros(np.shape(ndf_cpu))
+ diff_im = abs(ndf_cpu - ndf_gpu)
+ diff_im[diff_im > tolerance] = 1
+ self.assertLessEqual(diff_im.sum(), 1)
+
+
+ def test_Diff4th_CPU_vs_GPU(self):
+ filename = os.path.join("lena_gray_512.tif")
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+
+ Im = Im/255
+ perc = 0.05
+ u0 = Im + np.random.normal(loc = 0 ,
+ scale = perc * Im ,
+ size = np.shape(Im))
+ u_ref = Im + np.random.normal(loc = 0 ,
+ scale = 0.01 * Im ,
+ size = np.shape(Im))
+
+ # map the u0 u0->u0>0
+ # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1)
+ u0 = u0.astype('float32')
+ u_ref = u_ref.astype('float32')
+
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("___Anisotropic Diffusion 4th Order (2D)____")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+
+ # set parameters
+ pars = {'algorithm' : Diff4th, \
+ 'input' : u0,\
+ 'regularisation_parameter':3.5, \
+ 'edge_parameter':0.02,\
+ 'number_of_iterations' :500 ,\
+ 'time_marching_parameter':0.001
+ }
+
+ print ("#############Diff4th CPU####################")
+ start_time = timeit.default_timer()
+ diff4th_cpu = Diff4th(pars['input'],
+ pars['regularisation_parameter'],
+ pars['edge_parameter'],
+ pars['number_of_iterations'],
+ pars['time_marching_parameter'],'cpu')
+
+ rms = rmse(Im, diff4th_cpu)
+ pars['rmse'] = rms
+
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("##############Diff4th GPU##################")
+ start_time = timeit.default_timer()
+ try:
+ diff4th_gpu = Diff4th(pars['input'],
+ pars['regularisation_parameter'],
+ pars['edge_parameter'],
+ pars['number_of_iterations'],
+ pars['time_marching_parameter'], 'gpu')
+
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+ rms = rmse(Im, diff4th_gpu)
+ pars['rmse'] = rms
+ pars['algorithm'] = Diff4th
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("--------Compare the results--------")
+ tolerance = 1e-05
+ diff_im = np.zeros(np.shape(diff4th_cpu))
+ diff_im = abs(diff4th_cpu - diff4th_gpu)
+ diff_im[diff_im > tolerance] = 1
+ self.assertLessEqual(diff_im.sum() , 1)
+
+ def test_FDGdTV_CPU_vs_GPU(self):
+ filename = os.path.join("lena_gray_512.tif")
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+
+ Im = Im/255
+ perc = 0.05
+ u0 = Im + np.random.normal(loc = 0 ,
+ scale = perc * Im ,
+ size = np.shape(Im))
+ u_ref = Im + np.random.normal(loc = 0 ,
+ scale = 0.01 * Im ,
+ size = np.shape(Im))
+
+ # map the u0 u0->u0>0
+ # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1)
+ u0 = u0.astype('float32')
+ u_ref = u_ref.astype('float32')
+
+
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("____________FGP-dTV bench___________________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+
+ # set parameters
+ pars = {'algorithm' : FGP_dTV, \
+ 'input' : u0,\
+ 'refdata' : u_ref,\
+ 'regularisation_parameter':0.04, \
+ 'number_of_iterations' :1000 ,\
+ 'tolerance_constant':1e-07,\
+ 'eta_const':0.2,\
+ 'methodTV': 0 ,\
+ 'nonneg': 0 ,\
+ 'printingOut': 0
+ }
+
+ print ("#############FGP dTV CPU####################")
+ start_time = timeit.default_timer()
+ fgp_dtv_cpu = FGP_dTV(pars['input'],
+ pars['refdata'],
+ pars['regularisation_parameter'],
+ pars['number_of_iterations'],
+ pars['tolerance_constant'],
+ pars['eta_const'],
+ pars['methodTV'],
+ pars['nonneg'],
+ pars['printingOut'],'cpu')
+
+
+ rms = rmse(Im, fgp_dtv_cpu)
+ pars['rmse'] = rms
+
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("##############FGP dTV GPU##################")
+ start_time = timeit.default_timer()
+ try:
+ fgp_dtv_gpu = FGP_dTV(pars['input'],
+ pars['refdata'],
+ pars['regularisation_parameter'],
+ pars['number_of_iterations'],
+ pars['tolerance_constant'],
+ pars['eta_const'],
+ pars['methodTV'],
+ pars['nonneg'],
+ pars['printingOut'],'gpu')
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+ rms = rmse(Im, fgp_dtv_gpu)
+ pars['rmse'] = rms
+ pars['algorithm'] = FGP_dTV
+ txtstr = printParametersToString(pars)
+ txtstr += "%s = %.3fs" % ('elapsed time',timeit.default_timer() - start_time)
+ print (txtstr)
+ print ("--------Compare the results--------")
+ tolerance = 1e-05
+ diff_im = np.zeros(np.shape(fgp_dtv_cpu))
+ diff_im = abs(fgp_dtv_cpu - fgp_dtv_gpu)
+ diff_im[diff_im > tolerance] = 1
+ self.assertLessEqual(diff_im.sum(), 1)
+
+ def test_cpu_ROF_TV(self):
+ #filename = os.path.join(".." , ".." , ".." , "data" ,"testLena.npy")
+
+ filename = os.path.join("lena_gray_512.tif")
+
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+ Im = Im/255
+
+ """
+ # read noiseless image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+ """
+ tolerance = 1e-05
+ rms_rof_exp = 8.313131464999238e-05 #expected value for ROF model
+
+ # set parameters for ROF-TV
+ pars_rof_tv = {'algorithm': ROF_TV, \
+ 'input' : Im,\
+ 'regularisation_parameter':0.04,\
+ 'number_of_iterations': 50,\
+ 'time_marching_parameter': 0.00001
+ }
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("_________testing ROF-TV (2D, CPU)__________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ rof_cpu = ROF_TV(pars_rof_tv['input'],
+ pars_rof_tv['regularisation_parameter'],
+ pars_rof_tv['number_of_iterations'],
+ pars_rof_tv['time_marching_parameter'],'cpu')
+ rms_rof = rmse(Im, rof_cpu)
+
+ # now compare obtained rms with the expected value
+ self.assertLess(abs(rms_rof-rms_rof_exp) , tolerance)
+ def test_cpu_FGP_TV(self):
+ #filename = os.path.join(".." , ".." , ".." , "data" ,"testLena.npy")
+
+ filename = os.path.join("lena_gray_512.tif")
+
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+ Im = Im/255
+ """
+ # read noiseless image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+ """
+ tolerance = 1e-05
+ rms_fgp_exp = 0.019152347 #expected value for FGP model
+
+ pars_fgp_tv = {'algorithm' : FGP_TV, \
+ 'input' : Im,\
+ 'regularisation_parameter':0.04, \
+ 'number_of_iterations' :50 ,\
+ 'tolerance_constant':1e-06,\
+ 'methodTV': 0 ,\
+ 'nonneg': 0 ,\
+ 'printingOut': 0
+ }
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("_________testing FGP-TV (2D, CPU)__________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ fgp_cpu = FGP_TV(pars_fgp_tv['input'],
+ pars_fgp_tv['regularisation_parameter'],
+ pars_fgp_tv['number_of_iterations'],
+ pars_fgp_tv['tolerance_constant'],
+ pars_fgp_tv['methodTV'],
+ pars_fgp_tv['nonneg'],
+ pars_fgp_tv['printingOut'],'cpu')
+ rms_fgp = rmse(Im, fgp_cpu)
+ # now compare obtained rms with the expected value
+ self.assertLess(abs(rms_fgp-rms_fgp_exp) , tolerance)
+
+ def test_gpu_ROF(self):
+ #filename = os.path.join(".." , ".." , ".." , "data" ,"testLena.npy")
+ filename = os.path.join("lena_gray_512.tif")
+
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+ Im = Im/255
+
+ tolerance = 1e-05
+ rms_rof_exp = 8.313131464999238e-05 #expected value for ROF model
+
+ # set parameters for ROF-TV
+ pars_rof_tv = {'algorithm': ROF_TV, \
+ 'input' : Im,\
+ 'regularisation_parameter':0.04,\
+ 'number_of_iterations': 50,\
+ 'time_marching_parameter': 0.00001
+ }
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("_________testing ROF-TV (2D, GPU)__________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ try:
+ rof_gpu = ROF_TV(pars_rof_tv['input'],
+ pars_rof_tv['regularisation_parameter'],
+ pars_rof_tv['number_of_iterations'],
+ pars_rof_tv['time_marching_parameter'],'gpu')
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+
+ rms_rof = rmse(Im, rof_gpu)
+ # now compare obtained rms with the expected value
+ self.assertLess(abs(rms_rof-rms_rof_exp) , tolerance)
+
+ def test_gpu_FGP(self):
+ #filename = os.path.join(".." , ".." , ".." , "data" ,"testLena.npy")
+ filename = os.path.join("lena_gray_512.tif")
+
+ plt = TiffReader()
+ # read image
+ Im = plt.imread(filename)
+ Im = np.asarray(Im, dtype='float32')
+ Im = Im/255
+ tolerance = 1e-05
+
+ rms_fgp_exp = 0.019152347 #expected value for FGP model
+
+ # set parameters for FGP-TV
+ pars_fgp_tv = {'algorithm' : FGP_TV, \
+ 'input' : Im,\
+ 'regularisation_parameter':0.04, \
+ 'number_of_iterations' :50 ,\
+ 'tolerance_constant':1e-06,\
+ 'methodTV': 0 ,\
+ 'nonneg': 0 ,\
+ 'printingOut': 0
+ }
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ print ("_________testing FGP-TV (2D, GPU)__________")
+ print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+ try:
+ fgp_gpu = FGP_TV(pars_fgp_tv['input'],
+ pars_fgp_tv['regularisation_parameter'],
+ pars_fgp_tv['number_of_iterations'],
+ pars_fgp_tv['tolerance_constant'],
+ pars_fgp_tv['methodTV'],
+ pars_fgp_tv['nonneg'],
+ pars_fgp_tv['printingOut'],'gpu')
+ except ValueError as ve:
+ self.skipTest("Results not comparable. GPU computing error.")
+ rms_fgp = rmse(Im, fgp_gpu)
+ # now compare obtained rms with the expected value
+
+ self.assertLess(abs(rms_fgp-rms_fgp_exp) , tolerance)
+
+if __name__ == '__main__':
+ unittest.main()
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt new file mode 100644 index 0000000..bdcb8f4 --- /dev/null +++ b/src/CMakeLists.txt @@ -0,0 +1,19 @@ +# Copyright 2017 Edoardo Pasca +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +if (BUILD_MATLAB_WRAPPER) + add_subdirectory(Matlab) +endif() +if (BUILD_PYTHON_WRAPPER) + add_subdirectory(Python) +endif()
\ No newline at end of file diff --git a/src/Core/CCPiDefines.h b/src/Core/CCPiDefines.h new file mode 100644 index 0000000..d3038f9 --- /dev/null +++ b/src/Core/CCPiDefines.h @@ -0,0 +1,35 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Srikanth Nagella, Edoardo Pasca, Daniil Kazantsev + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at + +http://www.apache.org/licenses/LICENSE-2.0 + +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ +#ifndef CCPIDEFINES_H +#define CCPIDEFINES_H + +#if defined(_WIN32) || defined(__WIN32__) + #if defined(CCPiCore_EXPORTS) || defined(CCPiNexusWidget_EXPORTS) || defined(ContourTreeSegmentation_EXPORTS) || defined(ContourTree_EXPORTS)// add by CMake + #define CCPI_EXPORT __declspec(dllexport) + #define EXPIMP_TEMPLATE + #else + #define CCPI_EXPORT __declspec(dllimport) + #define EXPIMP_TEMPLATE extern + #endif /* CCPi_EXPORTS */ +#elif defined(linux) || defined(__linux) || defined(__APPLE__) + #define CCPI_EXPORT +#endif + +#endif diff --git a/src/Core/CMakeLists.txt b/src/Core/CMakeLists.txt new file mode 100644 index 0000000..b3c0dfb --- /dev/null +++ b/src/Core/CMakeLists.txt @@ -0,0 +1,151 @@ +# Copyright 2018 Edoardo Pasca +#cmake_minimum_required (VERSION 3.0) + +project(RGL_core) +#https://stackoverflow.com/questions/13298504/using-cmake-with-setup-py + +# The version number. + +set (CIL_VERSION $ENV{CIL_VERSION} CACHE INTERNAL "Core Imaging Library version" FORCE) + +# conda orchestrated build +message("CIL_VERSION ${CIL_VERSION}") +#include (GenerateExportHeader) + + +find_package(OpenMP) +if (OPENMP_FOUND) + set (CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${OpenMP_C_FLAGS}") + set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${OpenMP_CXX_FLAGS}") + set (CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} ${OpenMP_EXE_LINKER_FLAGS} ${OpenMP_CXX_FLAGS}") + set (CMAKE_SHARED_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} ${OpenMP_SHARED_LINKER_FLAGS} ${OpenMP_CXX_FLAGS}") + set (CMAKE_STATIC_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} ${OpenMP_STATIC_LINKER_FLAGS} ${OpenMP_CXX_FLAGS}") + +endif() + +## Build the regularisers package as a library +message("Creating Regularisers as a shared library") + +message("CMAKE_CXX_FLAGS ${CMAKE_CXX_FLAGS}") +message("CMAKE_C_FLAGS ${CMAKE_C_FLAGS}") +message("CMAKE_EXE_LINKER_FLAGS ${CMAKE_EXE_LINKER_FLAGS}") +message("CMAKE_SHARED_LINKER_FLAGS ${CMAKE_SHARED_LINKER_FLAGS}") +message("CMAKE_STATIC_LINKER_FLAGS ${CMAKE_STATIC_LINKER_FLAGS}") + +set(CMAKE_BUILD_TYPE "Release") + +if(WIN32) + set (FLAGS "/DWIN32 /EHsc /DCCPiCore_EXPORTS /openmp") + set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${FLAGS}") + set (CMAKE_C_FLAGS "${CMAKE_CXX_FLAGS} ${FLAGS}") + set (CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} /NODEFAULTLIB:MSVCRT.lib") + + set (EXTRA_LIBRARIES) + + message("library lib: ${LIBRARY_LIB}") + +elseif(UNIX) + set (FLAGS "-O2 -funsigned-char -Wall -Wl,--no-undefined -DCCPiReconstructionIterative_EXPORTS ") + set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${FLAGS}") + set (CMAKE_C_FLAGS "${CMAKE_CXX_FLAGS} ${FLAGS}") + + set (EXTRA_LIBRARIES + "gomp" + "m" + ) + +endif() +message("CMAKE_CXX_FLAGS ${CMAKE_CXX_FLAGS}") + +## Build the regularisers package as a library +message("Adding regularisers as a shared library") + +#set(CMAKE_C_COMPILER /apps/pgi/linux86-64/17.4/bin/pgcc) +#set(CMAKE_C_FLAGS "-acc -Minfo -ta=tesla:cc20 -openmp") +#set(CMAKE_C_FLAGS "-acc -Minfo -ta=multicore -openmp -fPIC") +add_library(cilreg SHARED + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/FGP_TV_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/SB_TV_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/TGV_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/Diffusion_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/Diffus4th_order_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/LLT_ROF_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/ROF_TV_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/FGP_dTV_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/TNV_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/Nonlocal_TV_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/PatchSelect_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/utils.c + ${CMAKE_CURRENT_SOURCE_DIR}/inpainters_CPU/Diffusion_Inpaint_core.c + ${CMAKE_CURRENT_SOURCE_DIR}/inpainters_CPU/NonlocalMarching_Inpaint_core.c + ) +target_link_libraries(cilreg ${EXTRA_LIBRARIES} ) +include_directories(cilreg PUBLIC + ${LIBRARY_INC}/include + ${CMAKE_CURRENT_SOURCE_DIR} + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_CPU/ + ${CMAKE_CURRENT_SOURCE_DIR}/inpainters_CPU/ ) + +## Install + +if (UNIX) +message ("I'd install into ${CMAKE_INSTALL_PREFIX}/lib") +install(TARGETS cilreg + LIBRARY DESTINATION lib + CONFIGURATIONS ${CMAKE_BUILD_TYPE} + ) +elseif(WIN32) +message ("I'd install into ${CMAKE_INSTALL_PREFIX} lib bin") + install(TARGETS cilreg + RUNTIME DESTINATION bin + ARCHIVE DESTINATION lib + CONFIGURATIONS ${CMAKE_BUILD_TYPE} + ) +endif() + + + +# GPU Regularisers +if (BUILD_CUDA) + find_package(CUDA) + if (CUDA_FOUND) + set(CUDA_NVCC_FLAGS "-Xcompiler -fPIC -shared -D_FORCE_INLINES") + message("CUDA FLAGS ${CUDA_NVCC_FLAGS}") + CUDA_ADD_LIBRARY(cilregcuda SHARED + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_GPU/TV_ROF_GPU_core.cu + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_GPU/TV_FGP_GPU_core.cu + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_GPU/TV_SB_GPU_core.cu + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_GPU/LLT_ROF_GPU_core.cu + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_GPU/TGV_GPU_core.cu + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_GPU/dTV_FGP_GPU_core.cu + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_GPU/NonlDiff_GPU_core.cu + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_GPU/Diffus_4thO_GPU_core.cu + ${CMAKE_CURRENT_SOURCE_DIR}/regularisers_GPU/PatchSelect_GPU_core.cu + ) + if (UNIX) + message ("I'd install into ${CMAKE_INSTALL_PREFIX}/lib") + install(TARGETS cilregcuda + LIBRARY DESTINATION lib + CONFIGURATIONS ${CMAKE_BUILD_TYPE} + ) + elseif(WIN32) + message ("I'd install into ${CMAKE_INSTALL_PREFIX} lib bin") + install(TARGETS cilregcuda + RUNTIME DESTINATION bin + ARCHIVE DESTINATION lib + CONFIGURATIONS ${CMAKE_BUILD_TYPE} + ) + endif() + else() + message("CUDA NOT FOUND") + endif() +endif() + +if (${BUILD_MATLAB_WRAPPER}) + if (WIN32) + install(TARGETS cilreg DESTINATION ${MATLAB_DEST}) + if (CUDA_FOUND) + install(TARGETS cilregcuda DESTINATION ${MATLAB_DEST}) + endif() + endif() +endif() diff --git a/src/Core/inpainters_CPU/Diffusion_Inpaint_core.c b/src/Core/inpainters_CPU/Diffusion_Inpaint_core.c new file mode 100644 index 0000000..08b168a --- /dev/null +++ b/src/Core/inpainters_CPU/Diffusion_Inpaint_core.c @@ -0,0 +1,322 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "Diffusion_Inpaint_core.h" +#include "utils.h" + +/*sign function*/ +int signNDF_inc(float x) { + return (x > 0) - (x < 0); +} + +/* C-OMP implementation of linear and nonlinear diffusion [1,2] for inpainting task (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Image/volume to inpaint + * 2. Mask of the same size as (1) in 'unsigned char' format (ones mark the region to inpaint, zeros belong to the data) + * 3. lambda - regularization parameter + * 4. Edge-preserving parameter (sigma), when sigma equals to zero nonlinear diffusion -> linear diffusion + * 5. Number of iterations, for explicit scheme >= 150 is recommended + * 6. tau - time-marching step for explicit scheme + * 7. Penalty type: 1 - Huber, 2 - Perona-Malik, 3 - Tukey Biweight + * + * Output: + * [1] Inpainted image/volume + * + * This function is based on the paper by + * [1] Perona, P. and Malik, J., 1990. Scale-space and edge detection using anisotropic diffusion. IEEE Transactions on pattern analysis and machine intelligence, 12(7), pp.629-639. + * [2] Black, M.J., Sapiro, G., Marimont, D.H. and Heeger, D., 1998. Robust anisotropic diffusion. IEEE Transactions on image processing, 7(3), pp.421-432. + */ + +float Diffusion_Inpaint_CPU_main(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int penaltytype, int dimX, int dimY, int dimZ) +{ + long i, pointsone; + float sigmaPar2; + sigmaPar2 = sigmaPar/sqrt(2.0f); + + /* copy into output */ + copyIm(Input, Output, (long)(dimX), (long)(dimY), (long)(dimZ)); + + pointsone = 0; + for (i=0; i<dimY*dimX*dimZ; i++) if (Mask[i] == 1) pointsone++; + + if (pointsone == 0) printf("%s \n", "Nothing to inpaint, zero mask!"); + else { + + if (dimZ == 1) { + /* running 2D diffusion iterations */ + for(i=0; i < iterationsNumb; i++) { + if (sigmaPar == 0.0f) LinearDiff_Inp_2D(Input, Mask, Output, lambdaPar, tau, (long)(dimX), (long)(dimY)); /* linear diffusion (heat equation) */ + else NonLinearDiff_Inp_2D(Input, Mask, Output, lambdaPar, sigmaPar2, tau, penaltytype, (long)(dimX), (long)(dimY)); /* nonlinear diffusion */ + } + } + else { + /* running 3D diffusion iterations */ + for(i=0; i < iterationsNumb; i++) { + if (sigmaPar == 0.0f) LinearDiff_Inp_3D(Input, Mask, Output, lambdaPar, tau, (long)(dimX), (long)(dimY), (long)(dimZ)); + else NonLinearDiff_Inp_3D(Input, Mask, Output, lambdaPar, sigmaPar2, tau, penaltytype, (long)(dimX), (long)(dimY), (long)(dimZ)); + } + } + } + return *Output; +} +/********************************************************************/ +/***************************2D Functions*****************************/ +/********************************************************************/ +/* linear diffusion (heat equation) */ +float LinearDiff_Inp_2D(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float tau, long dimX, long dimY) +{ + long i,j,i1,i2,j1,j2,index; + float e,w,n,s,e1,w1,n1,s1; + +#pragma omp parallel for shared(Input,Mask) private(index,i,j,i1,i2,j1,j2,e,w,n,s,e1,w1,n1,s1) + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = j*dimX+i; + + if (Mask[index] > 0) { + /*inpainting process*/ + e = Output[j*dimX+i1]; + w = Output[j*dimX+i2]; + n = Output[j1*dimX+i]; + s = Output[j2*dimX+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1) - (Output[index] - Input[index])); + } + }} + return *Output; +} + +/* nonlinear diffusion */ +float NonLinearDiff_Inp_2D(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, long dimX, long dimY) +{ + long i,j,i1,i2,j1,j2,index; + float e,w,n,s,e1,w1,n1,s1; + +#pragma omp parallel for shared(Input,Mask) private(index,i,j,i1,i2,j1,j2,e,w,n,s,e1,w1,n1,s1) + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = j*dimX+i; + + if (Mask[index] > 0) { + /*inpainting process*/ + e = Output[j*dimX+i1]; + w = Output[j*dimX+i2]; + n = Output[j1*dimX+i]; + s = Output[j2*dimX+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + + if (penaltytype == 1){ + /* Huber penalty */ + if (fabs(e1) > sigmaPar) e1 = signNDF_inc(e1); + else e1 = e1/sigmaPar; + + if (fabs(w1) > sigmaPar) w1 = signNDF_inc(w1); + else w1 = w1/sigmaPar; + + if (fabs(n1) > sigmaPar) n1 = signNDF_inc(n1); + else n1 = n1/sigmaPar; + + if (fabs(s1) > sigmaPar) s1 = signNDF_inc(s1); + else s1 = s1/sigmaPar; + } + else if (penaltytype == 2) { + /* Perona-Malik */ + e1 = (e1)/(1.0f + powf((e1/sigmaPar),2)); + w1 = (w1)/(1.0f + powf((w1/sigmaPar),2)); + n1 = (n1)/(1.0f + powf((n1/sigmaPar),2)); + s1 = (s1)/(1.0f + powf((s1/sigmaPar),2)); + } + else if (penaltytype == 3) { + /* Tukey Biweight */ + if (fabs(e1) <= sigmaPar) e1 = e1*powf((1.0f - powf((e1/sigmaPar),2)), 2); + else e1 = 0.0f; + if (fabs(w1) <= sigmaPar) w1 = w1*powf((1.0f - powf((w1/sigmaPar),2)), 2); + else w1 = 0.0f; + if (fabs(n1) <= sigmaPar) n1 = n1*powf((1.0f - powf((n1/sigmaPar),2)), 2); + else n1 = 0.0f; + if (fabs(s1) <= sigmaPar) s1 = s1*powf((1.0f - powf((s1/sigmaPar),2)), 2); + else s1 = 0.0f; + } + else { + printf("%s \n", "No penalty function selected! Use 1,2 or 3."); + break; + } + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1) - (Output[index] - Input[index])); + } + }} + return *Output; +} +/********************************************************************/ +/***************************3D Functions*****************************/ +/********************************************************************/ +/* linear diffusion (heat equation) */ +float LinearDiff_Inp_3D(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float tau, long dimX, long dimY, long dimZ) +{ + long i,j,k,i1,i2,j1,j2,k1,k2,index; + float e,w,n,s,u,d,e1,w1,n1,s1,u1,d1; + +#pragma omp parallel for shared(Input,Mask) private(index,i,j,i1,i2,j1,j2,e,w,n,s,e1,w1,n1,s1,k,k1,k2,u1,d1,u,d) +for(k=0; k<dimZ; k++) { + k1 = k+1; if (k1 == dimZ) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = (dimX*dimY)*k + j*dimX+i; + + if (Mask[index] > 0) { + /*inpainting process*/ + + e = Output[(dimX*dimY)*k + j*dimX+i1]; + w = Output[(dimX*dimY)*k + j*dimX+i2]; + n = Output[(dimX*dimY)*k + j1*dimX+i]; + s = Output[(dimX*dimY)*k + j2*dimX+i]; + u = Output[(dimX*dimY)*k1 + j*dimX+i]; + d = Output[(dimX*dimY)*k2 + j*dimX+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + u1 = u - Output[index]; + d1 = d - Output[index]; + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1 + u1 + d1) - (Output[index] - Input[index])); + } + }}} + return *Output; +} + +float NonLinearDiff_Inp_3D(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, long dimX, long dimY, long dimZ) +{ + long i,j,k,i1,i2,j1,j2,k1,k2,index; + float e,w,n,s,u,d,e1,w1,n1,s1,u1,d1; + +#pragma omp parallel for shared(Input,Mask) private(index,i,j,i1,i2,j1,j2,e,w,n,s,e1,w1,n1,s1,k,k1,k2,u1,d1,u,d) +for(k=0; k<dimZ; k++) { + k1 = k+1; if (k1 == dimZ) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = (dimX*dimY)*k + j*dimX+i; + + if (Mask[index] > 0) { + /*inpainting process*/ + e = Output[(dimX*dimY)*k + j*dimX+i1]; + w = Output[(dimX*dimY)*k + j*dimX+i2]; + n = Output[(dimX*dimY)*k + j1*dimX+i]; + s = Output[(dimX*dimY)*k + j2*dimX+i]; + u = Output[(dimX*dimY)*k1 + j*dimX+i]; + d = Output[(dimX*dimY)*k2 + j*dimX+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + u1 = u - Output[index]; + d1 = d - Output[index]; + + if (penaltytype == 1){ + /* Huber penalty */ + if (fabs(e1) > sigmaPar) e1 = signNDF_inc(e1); + else e1 = e1/sigmaPar; + + if (fabs(w1) > sigmaPar) w1 = signNDF_inc(w1); + else w1 = w1/sigmaPar; + + if (fabs(n1) > sigmaPar) n1 = signNDF_inc(n1); + else n1 = n1/sigmaPar; + + if (fabs(s1) > sigmaPar) s1 = signNDF_inc(s1); + else s1 = s1/sigmaPar; + + if (fabs(u1) > sigmaPar) u1 = signNDF_inc(u1); + else u1 = u1/sigmaPar; + + if (fabs(d1) > sigmaPar) d1 = signNDF_inc(d1); + else d1 = d1/sigmaPar; + } + else if (penaltytype == 2) { + /* Perona-Malik */ + e1 = (e1)/(1.0f + powf((e1/sigmaPar),2)); + w1 = (w1)/(1.0f + powf((w1/sigmaPar),2)); + n1 = (n1)/(1.0f + powf((n1/sigmaPar),2)); + s1 = (s1)/(1.0f + powf((s1/sigmaPar),2)); + u1 = (u1)/(1.0f + powf((u1/sigmaPar),2)); + d1 = (d1)/(1.0f + powf((d1/sigmaPar),2)); + } + else if (penaltytype == 3) { + /* Tukey Biweight */ + if (fabs(e1) <= sigmaPar) e1 = e1*powf((1.0f - powf((e1/sigmaPar),2)), 2); + else e1 = 0.0f; + if (fabs(w1) <= sigmaPar) w1 = w1*powf((1.0f - powf((w1/sigmaPar),2)), 2); + else w1 = 0.0f; + if (fabs(n1) <= sigmaPar) n1 = n1*powf((1.0f - powf((n1/sigmaPar),2)), 2); + else n1 = 0.0f; + if (fabs(s1) <= sigmaPar) s1 = s1*powf((1.0f - powf((s1/sigmaPar),2)), 2); + else s1 = 0.0f; + if (fabs(u1) <= sigmaPar) u1 = u1*powf((1.0f - powf((u1/sigmaPar),2)), 2); + else u1 = 0.0f; + if (fabs(d1) <= sigmaPar) d1 = d1*powf((1.0f - powf((d1/sigmaPar),2)), 2); + else d1 = 0.0f; + } + else { + printf("%s \n", "No penalty function selected! Use 1,2 or 3."); + break; + } + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1 + u1 + d1) - (Output[index] - Input[index])); + } + }}} + return *Output; +} diff --git a/src/Core/inpainters_CPU/Diffusion_Inpaint_core.h b/src/Core/inpainters_CPU/Diffusion_Inpaint_core.h new file mode 100644 index 0000000..a96fe79 --- /dev/null +++ b/src/Core/inpainters_CPU/Diffusion_Inpaint_core.h @@ -0,0 +1,61 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + + +/* C-OMP implementation of linear and nonlinear diffusion [1,2] for inpainting task (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Image/volume to inpaint + * 2. Mask of the same size as (1) in 'unsigned char' format (ones mark the region to inpaint, zeros belong to the data) + * 3. lambda - regularization parameter + * 4. Edge-preserving parameter (sigma), when sigma equals to zero nonlinear diffusion -> linear diffusion + * 5. Number of iterations, for explicit scheme >= 150 is recommended + * 6. tau - time-marching step for explicit scheme + * 7. Penalty type: 1 - Huber, 2 - Perona-Malik, 3 - Tukey Biweight + * + * Output: + * [1] Inpainted image/volume + * + * This function is based on the paper by + * [1] Perona, P. and Malik, J., 1990. Scale-space and edge detection using anisotropic diffusion. IEEE Transactions on pattern analysis and machine intelligence, 12(7), pp.629-639. + * [2] Black, M.J., Sapiro, G., Marimont, D.H. and Heeger, D., 1998. Robust anisotropic diffusion. IEEE Transactions on image processing, 7(3), pp.421-432. + */ + + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float Diffusion_Inpaint_CPU_main(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int penaltytype, int dimX, int dimY, int dimZ); + +CCPI_EXPORT float LinearDiff_Inp_2D(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float tau, long dimX, long dimY); +CCPI_EXPORT float NonLinearDiff_Inp_2D(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, long dimX, long dimY); +CCPI_EXPORT float LinearDiff_Inp_3D(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float tau, long dimX, long dimY, long dimZ); +CCPI_EXPORT float NonLinearDiff_Inp_3D(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, long dimX, long dimY, long dimZ); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/inpainters_CPU/NonlocalMarching_Inpaint_core.c b/src/Core/inpainters_CPU/NonlocalMarching_Inpaint_core.c new file mode 100644 index 0000000..b488ca4 --- /dev/null +++ b/src/Core/inpainters_CPU/NonlocalMarching_Inpaint_core.c @@ -0,0 +1,188 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "NonlocalMarching_Inpaint_core.h" +#include "utils.h" + + +/* C-OMP implementation of Nonlocal Vertical Marching inpainting method (2D case) + * The method is heuristic but computationally efficent (especially for larger images). + * It developed specifically to smoothly inpaint horizontal or inclined missing data regions in sinograms + * The method WILL not work satisfactory if you have lengthy vertical stripes of missing data + * + * Input: + * 1. 2D image or sinogram with horizontal or inclined regions of missing data + * 2. Mask of the same size as A in 'unsigned char' format (ones mark the region to inpaint, zeros belong to the data) + * 3. Linear increment to increase searching window size in iterations, values from 1-3 is a good choice + * + * Output: + * 1. Inpainted image or a sinogram + * 2. updated mask + * + * Reference: D. Kazantsev (paper in preparation) + */ + +float NonlocalMarching_Inpaint_main(float *Input, unsigned char *M, float *Output, unsigned char *M_upd, int SW_increment, int iterationsNumb, int trigger, int dimX, int dimY, int dimZ) +{ + int i, j, i_m, j_m, counter, iter, iterations_number, W_fullsize, switchmask, switchcurr, counterElements; + float *Gauss_weights; + + /* copying M to M_upd */ + copyIm_unchar(M, M_upd, dimX, dimY, 1); + + /* Copying the image */ + copyIm(Input, Output, dimX, dimY, 1); + + /* Find how many inpainting iterations (equal to the number of ones) required based on a mask */ + if (iterationsNumb == 0) { + iterations_number = 0; + for (i=0; i<dimY*dimX; i++) { + if (M[i] == 1) iterations_number++; + } + if ((int)(iterations_number/dimY) > dimX) iterations_number = dimX; + } + else iterations_number = iterationsNumb; + + if (iterations_number == 0) printf("%s \n", "Nothing to inpaint, zero mask!"); + else { + + printf("%s %i \n", "Max iteration number equals to:", iterations_number); + + /* Inpainting iterations run here*/ + int W_halfsize = 1; + for(iter=0; iter < iterations_number; iter++) { + + //if (mod (iter, 2) == 0) {W_halfsize += 1;} + // printf("%i \n", W_halfsize); + + /* pre-calculation of Gaussian distance weights */ + W_fullsize = (int)(2*W_halfsize + 1); /*full size of similarity window */ + Gauss_weights = (float*)calloc(W_fullsize*W_fullsize,sizeof(float )); + counter = 0; + for(i_m=-W_halfsize; i_m<=W_halfsize; i_m++) { + for(j_m=-W_halfsize; j_m<=W_halfsize; j_m++) { + Gauss_weights[counter] = exp(-(pow((i_m), 2) + pow((j_m), 2))/(2*W_halfsize*W_halfsize)); + counter++; + } + } + + if (trigger == 0) { + /*Matlab*/ +#pragma omp parallel for shared(Output, M_upd, Gauss_weights) private(i, j, switchmask, switchcurr) + for(j=0; j<dimY; j++) { + switchmask = 0; + for(i=0; i<dimX; i++) { + switchcurr = 0; + if ((M_upd[j*dimX + i] == 1) && (switchmask == 0)) { + /* perform inpainting of the current pixel */ + inpaint_func(Output, M_upd, Gauss_weights, i, j, dimX, dimY, W_halfsize, W_fullsize); + /* add value to the mask*/ + M_upd[j*dimX + i] = 0; + switchmask = 1; switchcurr = 1; + } + if ((M_upd[j*dimX + i] == 0) && (switchmask == 1) && (switchcurr == 0)) { + /* perform inpainting of the previous (i-1) pixel */ + inpaint_func(Output, M_upd, Gauss_weights, i-1, j, dimX, dimY, W_halfsize, W_fullsize); + /* add value to the mask*/ + M_upd[(j)*dimX + i-1] = 0; + switchmask = 0; + } + } + } + } + else { + /*Python*/ + /* find a point in the mask to inpaint */ +#pragma omp parallel for shared(Output, M_upd, Gauss_weights) private(i, j, switchmask, switchcurr) + for(i=0; i<dimX; i++) { + switchmask = 0; + for(j=0; j<dimY; j++) { + switchcurr = 0; + if ((M_upd[j*dimX + i] == 1) && (switchmask == 0)) { + /* perform inpainting of the current pixel */ + inpaint_func(Output, M_upd, Gauss_weights, i, j, dimX, dimY, W_halfsize, W_fullsize); + /* add value to the mask*/ + M_upd[j*dimX + i] = 0; + switchmask = 1; switchcurr = 1; + } + if ((M_upd[j*dimX + i] == 0) && (switchmask == 1) && (switchcurr == 0)) { + /* perform inpainting of the previous (j-1) pixel */ + inpaint_func(Output, M_upd, Gauss_weights, i, j-1, dimX, dimY, W_halfsize, W_fullsize); + /* add value to the mask*/ + M_upd[(j-1)*dimX + i] = 0; + switchmask = 0; + } + } + } + } + free(Gauss_weights); + + /* check if possible to terminate iterations earlier */ + counterElements = 0; + for(i=0; i<dimX*dimY; i++) if (M_upd[i] == 0) counterElements++; + + if (counterElements == dimX*dimY) { + printf("%s \n", "Padding completed!"); + break; + } + W_halfsize += SW_increment; + } + printf("%s %i \n", "Iterations stopped at:", iter); + } + return *Output; +} + +float inpaint_func(float *U, unsigned char *M_upd, float *Gauss_weights, int i, int j, int dimX, int dimY, int W_halfsize, int W_fullsize) +{ + int i1, j1, i_m, j_m, counter; + float sum_val, sumweight; + + /*method 1: inpainting based on Euclidian weights */ + sumweight = 0.0f; + counter = 0; sum_val = 0.0f; + for(i_m=-W_halfsize; i_m<=W_halfsize; i_m++) { + i1 = i+i_m; + for(j_m=-W_halfsize; j_m<=W_halfsize; j_m++) { + j1 = j+j_m; + if (((i1 >= 0) && (i1 < dimX)) && ((j1 >= 0) && (j1 < dimY))) { + if (M_upd[j1*dimX + i1] == 0) { + sumweight += Gauss_weights[counter]; + } + } + counter++; + } + } + counter = 0; sum_val = 0.0f; + for(i_m=-W_halfsize; i_m<=W_halfsize; i_m++) { + i1 = i+i_m; + for(j_m=-W_halfsize; j_m<=W_halfsize; j_m++) { + j1 = j+j_m; + if (((i1 >= 0) && (i1 < dimX)) && ((j1 >= 0) && (j1 < dimY))) { + if ((M_upd[j1*dimX + i1] == 0) && (sumweight != 0.0f)) { + /* we have data so add it with Euc weight */ + sum_val += (Gauss_weights[counter]/sumweight)*U[j1*dimX + i1]; + } + } + counter++; + } + } + U[j*dimX + i] = sum_val; + return *U; +} + diff --git a/src/Core/inpainters_CPU/NonlocalMarching_Inpaint_core.h b/src/Core/inpainters_CPU/NonlocalMarching_Inpaint_core.h new file mode 100644 index 0000000..0f99ed4 --- /dev/null +++ b/src/Core/inpainters_CPU/NonlocalMarching_Inpaint_core.h @@ -0,0 +1,54 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + + +/* C-OMP implementation of Nonlocal Vertical Marching inpainting method (2D case) + * The method is heuristic but computationally efficent (especially for larger images). + * It developed specifically to smoothly inpaint horizontal or inclined missing data regions in sinograms + * The method WILL not work satisfactory if you have lengthy vertical stripes of missing data + * + * Inputs: + * 1. 2D image or sinogram with horizontal or inclined regions of missing data + * 2. Mask of the same size as A in 'unsigned char' format (ones mark the region to inpaint, zeros belong to the data) + * 3. Linear increment to increase searching window size in iterations, values from 1-3 is a good choice + + * Output: + * 1. Inpainted image or a sinogram + * 2. updated mask + * + * Reference: TBA + */ + + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float NonlocalMarching_Inpaint_main(float *Input, unsigned char *M, float *Output, unsigned char *M_upd, int SW_increment, int iterationsNumb, int trigger, int dimX, int dimY, int dimZ); +CCPI_EXPORT float inpaint_func(float *U, unsigned char *M_upd, float *Gauss_weights, int i, int j, int dimX, int dimY, int W_halfsize, int W_fullsize); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/Diffus4th_order_core.c b/src/Core/regularisers_CPU/Diffus4th_order_core.c new file mode 100644 index 0000000..01f4f64 --- /dev/null +++ b/src/Core/regularisers_CPU/Diffus4th_order_core.c @@ -0,0 +1,250 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "Diffus4th_order_core.h" +#include "utils.h" + +#define EPS 1.0e-7 + +/* C-OMP implementation of fourth-order diffusion scheme [1] for piecewise-smooth recovery (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambda - regularization parameter + * 3. Edge-preserving parameter (sigma) + * 4. Number of iterations, for explicit scheme >= 150 is recommended + * 5. tau - time-marching step for the explicit scheme + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Hajiaboli, M.R., 2011. An anisotropic fourth-order diffusion filter for image noise removal. International Journal of Computer Vision, 92(2), pp.177-191. + */ + +float Diffus4th_CPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int dimX, int dimY, int dimZ) +{ + int i,DimTotal; + float sigmaPar2; + float *W_Lapl=NULL; + sigmaPar2 = sigmaPar*sigmaPar; + DimTotal = dimX*dimY*dimZ; + + W_Lapl = calloc(DimTotal, sizeof(float)); + + /* copy into output */ + copyIm(Input, Output, (long)(dimX), (long)(dimY), (long)(dimZ)); + + if (dimZ == 1) { + /* running 2D diffusion iterations */ + for(i=0; i < iterationsNumb; i++) { + /* Calculating weighted Laplacian */ + Weighted_Laplc2D(W_Lapl, Output, sigmaPar2, dimX, dimY); + /* Perform iteration step */ + Diffusion_update_step2D(Output, Input, W_Lapl, lambdaPar, sigmaPar2, tau, (long)(dimX), (long)(dimY)); + } + } + else { + /* running 3D diffusion iterations */ + for(i=0; i < iterationsNumb; i++) { + /* Calculating weighted Laplacian */ + Weighted_Laplc3D(W_Lapl, Output, sigmaPar2, dimX, dimY, dimZ); + /* Perform iteration step */ + Diffusion_update_step3D(Output, Input, W_Lapl, lambdaPar, sigmaPar2, tau, (long)(dimX), (long)(dimY), (long)(dimZ)); + } + } + free(W_Lapl); + return *Output; +} +/********************************************************************/ +/***************************2D Functions*****************************/ +/********************************************************************/ +float Weighted_Laplc2D(float *W_Lapl, float *U0, float sigma, long dimX, long dimY) +{ + long i,j,i1,i2,j1,j2,index; + float gradX, gradX_sq, gradY, gradY_sq, gradXX, gradYY, gradXY, xy_2, denom, V_norm, V_orth, c, c_sq; + + #pragma omp parallel for shared(W_Lapl) private(i,j,i1,i2,j1,j2,index,gradX, gradX_sq, gradY, gradY_sq, gradXX, gradYY, gradXY, xy_2, denom, V_norm, V_orth, c, c_sq) + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + + index = j*dimX+i; + + gradX = 0.5f*(U0[j*dimX+i2] - U0[j*dimX+i1]); + gradX_sq = pow(gradX,2); + + gradY = 0.5f*(U0[j2*dimX+i] - U0[j1*dimX+i]); + gradY_sq = pow(gradY,2); + + gradXX = U0[j*dimX+i2] + U0[j*dimX+i1] - 2*U0[index]; + gradYY = U0[j2*dimX+i] + U0[j1*dimX+i] - 2*U0[index]; + + gradXY = 0.25f*(U0[j2*dimX+i2] + U0[j1*dimX+i1] - U0[j1*dimX+i2] - U0[j2*dimX+i1]); + xy_2 = 2.0f*gradX*gradY*gradXY; + + denom = gradX_sq + gradY_sq; + + if (denom <= EPS) { + V_norm = (gradXX*gradX_sq + xy_2 + gradYY*gradY_sq)/EPS; + V_orth = (gradXX*gradY_sq - xy_2 + gradYY*gradX_sq)/EPS; + } + else { + V_norm = (gradXX*gradX_sq + xy_2 + gradYY*gradY_sq)/denom; + V_orth = (gradXX*gradY_sq - xy_2 + gradYY*gradX_sq)/denom; + } + + c = 1.0f/(1.0f + denom/sigma); + c_sq = c*c; + + W_Lapl[index] = c_sq*V_norm + c*V_orth; + } + } + return *W_Lapl; +} + +float Diffusion_update_step2D(float *Output, float *Input, float *W_Lapl, float lambdaPar, float sigmaPar2, float tau, long dimX, long dimY) +{ + long i,j,i1,i2,j1,j2,index; + float gradXXc, gradYYc; + + #pragma omp parallel for shared(Output, Input, W_Lapl) private(i,j,i1,i2,j1,j2,index,gradXXc,gradYYc) + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = j*dimX+i; + + gradXXc = W_Lapl[j*dimX+i2] + W_Lapl[j*dimX+i1] - 2*W_Lapl[index]; + gradYYc = W_Lapl[j2*dimX+i] + W_Lapl[j1*dimX+i] - 2*W_Lapl[index]; + + Output[index] += tau*(-lambdaPar*(gradXXc + gradYYc) - (Output[index] - Input[index])); + } + } + return *Output; +} +/********************************************************************/ +/***************************3D Functions*****************************/ +/********************************************************************/ +float Weighted_Laplc3D(float *W_Lapl, float *U0, float sigma, long dimX, long dimY, long dimZ) +{ + long i,j,k,i1,i2,j1,j2,k1,k2,index; + float gradX, gradX_sq, gradY, gradY_sq, gradXX, gradYY, gradXY, xy_2, denom, V_norm, V_orth, c, c_sq, gradZ, gradZ_sq, gradZZ, gradXZ, gradYZ, xyz_1, xyz_2; + + #pragma omp parallel for shared(W_Lapl) private(i,j,k,i1,i2,j1,j2,k1,k2,index,gradX, gradX_sq, gradY, gradY_sq, gradXX, gradYY, gradXY, xy_2, denom, V_norm, V_orth, c, c_sq, gradZ, gradZ_sq, gradZZ, gradXZ, gradYZ, xyz_1, xyz_2) + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + + for(k=0; k<dimZ; k++) { + /* symmetric boundary conditions */ + k1 = k+1; if (k1 == dimZ) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + + index = (dimX*dimY)*k + j*dimX+i; + + gradX = 0.5f*(U0[(dimX*dimY)*k + j*dimX+i2] - U0[(dimX*dimY)*k + j*dimX+i1]); + gradX_sq = pow(gradX,2); + + gradY = 0.5f*(U0[(dimX*dimY)*k + j2*dimX+i] - U0[(dimX*dimY)*k + j1*dimX+i]); + gradY_sq = pow(gradY,2); + + gradZ = 0.5f*(U0[(dimX*dimY)*k2 + j*dimX+i] - U0[(dimX*dimY)*k1 + j*dimX+i]); + gradZ_sq = pow(gradZ,2); + + gradXX = U0[(dimX*dimY)*k + j*dimX+i2] + U0[(dimX*dimY)*k + j*dimX+i1] - 2*U0[index]; + gradYY = U0[(dimX*dimY)*k + j2*dimX+i] + U0[(dimX*dimY)*k + j1*dimX+i] - 2*U0[index]; + gradZZ = U0[(dimX*dimY)*k2 + j*dimX+i] + U0[(dimX*dimY)*k1 + j*dimX+i] - 2*U0[index]; + + gradXY = 0.25f*(U0[(dimX*dimY)*k + j2*dimX+i2] + U0[(dimX*dimY)*k + j1*dimX+i1] - U0[(dimX*dimY)*k + j1*dimX+i2] - U0[(dimX*dimY)*k + j2*dimX+i1]); + gradXZ = 0.25f*(U0[(dimX*dimY)*k2 + j*dimX+i2] - U0[(dimX*dimY)*k2+j*dimX+i1] - U0[(dimX*dimY)*k1+j*dimX+i2] + U0[(dimX*dimY)*k1+j*dimX+i1]); + gradYZ = 0.25f*(U0[(dimX*dimY)*k2 +j2*dimX+i] - U0[(dimX*dimY)*k2+j1*dimX+i] - U0[(dimX*dimY)*k1+j2*dimX+i] + U0[(dimX*dimY)*k1+j1*dimX+i]); + + xy_2 = 2.0f*gradX*gradY*gradXY; + xyz_1 = 2.0f*gradX*gradZ*gradXZ; + xyz_2 = 2.0f*gradY*gradZ*gradYZ; + + denom = gradX_sq + gradY_sq + gradZ_sq; + + if (denom <= EPS) { + V_norm = (gradXX*gradX_sq + gradYY*gradY_sq + gradZZ*gradZ_sq + xy_2 + xyz_1 + xyz_2)/EPS; + V_orth = ((gradY_sq + gradZ_sq)*gradXX + (gradX_sq + gradZ_sq)*gradYY + (gradX_sq + gradY_sq)*gradZZ - xy_2 - xyz_1 - xyz_2)/EPS; + } + else { + V_norm = (gradXX*gradX_sq + gradYY*gradY_sq + gradZZ*gradZ_sq + xy_2 + xyz_1 + xyz_2)/denom; + V_orth = ((gradY_sq + gradZ_sq)*gradXX + (gradX_sq + gradZ_sq)*gradYY + (gradX_sq + gradY_sq)*gradZZ - xy_2 - xyz_1 - xyz_2)/denom; + } + + c = 1.0f/(1.0f + denom/sigma); + c_sq = c*c; + + W_Lapl[index] = c_sq*V_norm + c*V_orth; + } + } + } + return *W_Lapl; +} + +float Diffusion_update_step3D(float *Output, float *Input, float *W_Lapl, float lambdaPar, float sigmaPar2, float tau, long dimX, long dimY, long dimZ) +{ + long i,j,i1,i2,j1,j2,index,k,k1,k2; + float gradXXc, gradYYc, gradZZc; + + #pragma omp parallel for shared(Output, Input, W_Lapl) private(i,j,i1,i2,j1,j2,k,k1,k2,index,gradXXc,gradYYc,gradZZc) + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + + for(k=0; k<dimZ; k++) { + /* symmetric boundary conditions */ + k1 = k+1; if (k1 == dimZ) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + + index = (dimX*dimY)*k + j*dimX+i; + + gradXXc = W_Lapl[(dimX*dimY)*k + j*dimX+i2] + W_Lapl[(dimX*dimY)*k + j*dimX+i1] - 2*W_Lapl[index]; + gradYYc = W_Lapl[(dimX*dimY)*k + j2*dimX+i] + W_Lapl[(dimX*dimY)*k + j1*dimX+i] - 2*W_Lapl[index]; + gradZZc = W_Lapl[(dimX*dimY)*k2 + j*dimX+i] + W_Lapl[(dimX*dimY)*k1 + j*dimX+i] - 2*W_Lapl[index]; + + Output[index] += tau*(-lambdaPar*(gradXXc + gradYYc + gradZZc) - (Output[index] - Input[index])); + } + } + } + return *Output; +} diff --git a/src/Core/regularisers_CPU/Diffus4th_order_core.h b/src/Core/regularisers_CPU/Diffus4th_order_core.h new file mode 100644 index 0000000..d81afcb --- /dev/null +++ b/src/Core/regularisers_CPU/Diffus4th_order_core.h @@ -0,0 +1,55 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + +/* C-OMP implementation of fourth-order diffusion scheme [1] for piecewise-smooth recovery (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambda - regularization parameter + * 3. Edge-preserving parameter (sigma) + * 4. Number of iterations, for explicit scheme >= 150 is recommended + * 5. tau - time-marching step for explicit scheme + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Hajiaboli, M.R., 2011. An anisotropic fourth-order diffusion filter for image noise removal. International Journal of Computer Vision, 92(2), pp.177-191. + */ + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float Diffus4th_CPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int dimX, int dimY, int dimZ); +CCPI_EXPORT float Weighted_Laplc2D(float *W_Lapl, float *U0, float sigma, long dimX, long dimY); +CCPI_EXPORT float Diffusion_update_step2D(float *Output, float *Input, float *W_Lapl, float lambdaPar, float sigmaPar2, float tau, long dimX, long dimY); +CCPI_EXPORT float Weighted_Laplc3D(float *W_Lapl, float *U0, float sigma, long dimX, long dimY, long dimZ); +CCPI_EXPORT float Diffusion_update_step3D(float *Output, float *Input, float *W_Lapl, float lambdaPar, float sigmaPar2, float tau, long dimX, long dimY, long dimZ); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/Diffusion_core.c b/src/Core/regularisers_CPU/Diffusion_core.c new file mode 100644 index 0000000..b765796 --- /dev/null +++ b/src/Core/regularisers_CPU/Diffusion_core.c @@ -0,0 +1,307 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "Diffusion_core.h" +#include "utils.h" + +#define EPS 1.0e-5 +#define MAX(x, y) (((x) > (y)) ? (x) : (y)) +#define MIN(x, y) (((x) < (y)) ? (x) : (y)) + +/*sign function*/ +int signNDFc(float x) { + return (x > 0) - (x < 0); +} + +/* C-OMP implementation of linear and nonlinear diffusion with the regularisation model [1,2] (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambda - regularization parameter + * 3. Edge-preserving parameter (sigma), when sigma equals to zero nonlinear diffusion -> linear diffusion + * 4. Number of iterations, for explicit scheme >= 150 is recommended + * 5. tau - time-marching step for explicit scheme + * 6. Penalty type: 1 - Huber, 2 - Perona-Malik, 3 - Tukey Biweight + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Perona, P. and Malik, J., 1990. Scale-space and edge detection using anisotropic diffusion. IEEE Transactions on pattern analysis and machine intelligence, 12(7), pp.629-639. + * [2] Black, M.J., Sapiro, G., Marimont, D.H. and Heeger, D., 1998. Robust anisotropic diffusion. IEEE Transactions on image processing, 7(3), pp.421-432. + */ + +float Diffusion_CPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int penaltytype, int dimX, int dimY, int dimZ) +{ + int i; + float sigmaPar2; + sigmaPar2 = sigmaPar/sqrt(2.0f); + + /* copy into output */ + copyIm(Input, Output, (long)(dimX), (long)(dimY), (long)(dimZ)); + + if (dimZ == 1) { + /* running 2D diffusion iterations */ + for(i=0; i < iterationsNumb; i++) { + if (sigmaPar == 0.0f) LinearDiff2D(Input, Output, lambdaPar, tau, (long)(dimX), (long)(dimY)); /* linear diffusion (heat equation) */ + else NonLinearDiff2D(Input, Output, lambdaPar, sigmaPar2, tau, penaltytype, (long)(dimX), (long)(dimY)); /* nonlinear diffusion */ + } + } + else { + /* running 3D diffusion iterations */ + for(i=0; i < iterationsNumb; i++) { + if (sigmaPar == 0.0f) LinearDiff3D(Input, Output, lambdaPar, tau, (long)(dimX), (long)(dimY), (long)(dimZ)); + else NonLinearDiff3D(Input, Output, lambdaPar, sigmaPar2, tau, penaltytype, (long)(dimX), (long)(dimY), (long)(dimZ)); + } + } + return *Output; +} + + +/********************************************************************/ +/***************************2D Functions*****************************/ +/********************************************************************/ +/* linear diffusion (heat equation) */ +float LinearDiff2D(float *Input, float *Output, float lambdaPar, float tau, long dimX, long dimY) +{ + long i,j,i1,i2,j1,j2,index; + float e,w,n,s,e1,w1,n1,s1; + +#pragma omp parallel for shared(Input) private(index,i,j,i1,i2,j1,j2,e,w,n,s,e1,w1,n1,s1) + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = j*dimX+i; + + e = Output[j*dimX+i1]; + w = Output[j*dimX+i2]; + n = Output[j1*dimX+i]; + s = Output[j2*dimX+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1) - (Output[index] - Input[index])); + }} + return *Output; +} + +/* nonlinear diffusion */ +float NonLinearDiff2D(float *Input, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, long dimX, long dimY) +{ + long i,j,i1,i2,j1,j2,index; + float e,w,n,s,e1,w1,n1,s1; + +#pragma omp parallel for shared(Input) private(index,i,j,i1,i2,j1,j2,e,w,n,s,e1,w1,n1,s1) + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = j*dimX+i; + + e = Output[j*dimX+i1]; + w = Output[j*dimX+i2]; + n = Output[j1*dimX+i]; + s = Output[j2*dimX+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + + if (penaltytype == 1){ + /* Huber penalty */ + if (fabs(e1) > sigmaPar) e1 = signNDFc(e1); + else e1 = e1/sigmaPar; + + if (fabs(w1) > sigmaPar) w1 = signNDFc(w1); + else w1 = w1/sigmaPar; + + if (fabs(n1) > sigmaPar) n1 = signNDFc(n1); + else n1 = n1/sigmaPar; + + if (fabs(s1) > sigmaPar) s1 = signNDFc(s1); + else s1 = s1/sigmaPar; + } + else if (penaltytype == 2) { + /* Perona-Malik */ + e1 = (e1)/(1.0f + powf((e1/sigmaPar),2)); + w1 = (w1)/(1.0f + powf((w1/sigmaPar),2)); + n1 = (n1)/(1.0f + powf((n1/sigmaPar),2)); + s1 = (s1)/(1.0f + powf((s1/sigmaPar),2)); + } + else if (penaltytype == 3) { + /* Tukey Biweight */ + if (fabs(e1) <= sigmaPar) e1 = e1*powf((1.0f - powf((e1/sigmaPar),2)), 2); + else e1 = 0.0f; + if (fabs(w1) <= sigmaPar) w1 = w1*powf((1.0f - powf((w1/sigmaPar),2)), 2); + else w1 = 0.0f; + if (fabs(n1) <= sigmaPar) n1 = n1*powf((1.0f - powf((n1/sigmaPar),2)), 2); + else n1 = 0.0f; + if (fabs(s1) <= sigmaPar) s1 = s1*powf((1.0f - powf((s1/sigmaPar),2)), 2); + else s1 = 0.0f; + } + else { + printf("%s \n", "No penalty function selected! Use 1,2 or 3."); + break; + } + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1) - (Output[index] - Input[index])); + }} + return *Output; +} +/********************************************************************/ +/***************************3D Functions*****************************/ +/********************************************************************/ +/* linear diffusion (heat equation) */ +float LinearDiff3D(float *Input, float *Output, float lambdaPar, float tau, long dimX, long dimY, long dimZ) +{ + long i,j,k,i1,i2,j1,j2,k1,k2,index; + float e,w,n,s,u,d,e1,w1,n1,s1,u1,d1; + +#pragma omp parallel for shared(Input) private(index,i,j,i1,i2,j1,j2,e,w,n,s,e1,w1,n1,s1,k,k1,k2,u1,d1,u,d) +for(k=0; k<dimZ; k++) { + k1 = k+1; if (k1 == dimZ) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = (dimX*dimY)*k + j*dimX+i; + + e = Output[(dimX*dimY)*k + j*dimX+i1]; + w = Output[(dimX*dimY)*k + j*dimX+i2]; + n = Output[(dimX*dimY)*k + j1*dimX+i]; + s = Output[(dimX*dimY)*k + j2*dimX+i]; + u = Output[(dimX*dimY)*k1 + j*dimX+i]; + d = Output[(dimX*dimY)*k2 + j*dimX+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + u1 = u - Output[index]; + d1 = d - Output[index]; + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1 + u1 + d1) - (Output[index] - Input[index])); + }}} + return *Output; +} + +float NonLinearDiff3D(float *Input, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, long dimX, long dimY, long dimZ) +{ + long i,j,k,i1,i2,j1,j2,k1,k2,index; + float e,w,n,s,u,d,e1,w1,n1,s1,u1,d1; + +#pragma omp parallel for shared(Input) private(index,i,j,i1,i2,j1,j2,e,w,n,s,e1,w1,n1,s1,k,k1,k2,u1,d1,u,d) +for(k=0; k<dimZ; k++) { + k1 = k+1; if (k1 == dimZ) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = (dimX*dimY)*k + j*dimX+i; + + e = Output[(dimX*dimY)*k + j*dimX+i1]; + w = Output[(dimX*dimY)*k + j*dimX+i2]; + n = Output[(dimX*dimY)*k + j1*dimX+i]; + s = Output[(dimX*dimY)*k + j2*dimX+i]; + u = Output[(dimX*dimY)*k1 + j*dimX+i]; + d = Output[(dimX*dimY)*k2 + j*dimX+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + u1 = u - Output[index]; + d1 = d - Output[index]; + + if (penaltytype == 1){ + /* Huber penalty */ + if (fabs(e1) > sigmaPar) e1 = signNDFc(e1); + else e1 = e1/sigmaPar; + + if (fabs(w1) > sigmaPar) w1 = signNDFc(w1); + else w1 = w1/sigmaPar; + + if (fabs(n1) > sigmaPar) n1 = signNDFc(n1); + else n1 = n1/sigmaPar; + + if (fabs(s1) > sigmaPar) s1 = signNDFc(s1); + else s1 = s1/sigmaPar; + + if (fabs(u1) > sigmaPar) u1 = signNDFc(u1); + else u1 = u1/sigmaPar; + + if (fabs(d1) > sigmaPar) d1 = signNDFc(d1); + else d1 = d1/sigmaPar; + } + else if (penaltytype == 2) { + /* Perona-Malik */ + e1 = (e1)/(1.0f + powf((e1/sigmaPar),2)); + w1 = (w1)/(1.0f + powf((w1/sigmaPar),2)); + n1 = (n1)/(1.0f + powf((n1/sigmaPar),2)); + s1 = (s1)/(1.0f + powf((s1/sigmaPar),2)); + u1 = (u1)/(1.0f + powf((u1/sigmaPar),2)); + d1 = (d1)/(1.0f + powf((d1/sigmaPar),2)); + } + else if (penaltytype == 3) { + /* Tukey Biweight */ + if (fabs(e1) <= sigmaPar) e1 = e1*powf((1.0f - powf((e1/sigmaPar),2)), 2); + else e1 = 0.0f; + if (fabs(w1) <= sigmaPar) w1 = w1*powf((1.0f - powf((w1/sigmaPar),2)), 2); + else w1 = 0.0f; + if (fabs(n1) <= sigmaPar) n1 = n1*powf((1.0f - powf((n1/sigmaPar),2)), 2); + else n1 = 0.0f; + if (fabs(s1) <= sigmaPar) s1 = s1*powf((1.0f - powf((s1/sigmaPar),2)), 2); + else s1 = 0.0f; + if (fabs(u1) <= sigmaPar) u1 = u1*powf((1.0f - powf((u1/sigmaPar),2)), 2); + else u1 = 0.0f; + if (fabs(d1) <= sigmaPar) d1 = d1*powf((1.0f - powf((d1/sigmaPar),2)), 2); + else d1 = 0.0f; + } + else { + printf("%s \n", "No penalty function selected! Use 1,2 or 3."); + break; + } + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1 + u1 + d1) - (Output[index] - Input[index])); + }}} + return *Output; +} diff --git a/src/Core/regularisers_CPU/Diffusion_core.h b/src/Core/regularisers_CPU/Diffusion_core.h new file mode 100644 index 0000000..cc36dad --- /dev/null +++ b/src/Core/regularisers_CPU/Diffusion_core.h @@ -0,0 +1,59 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + + +/* C-OMP implementation of linear and nonlinear diffusion with the regularisation model [1,2] (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambda - regularization parameter + * 3. Edge-preserving parameter (sigma), when sigma equals to zero nonlinear diffusion -> linear diffusion + * 4. Number of iterations, for explicit scheme >= 150 is recommended + * 5. tau - time-marching step for explicit scheme + * 6. Penalty type: 1 - Huber, 2 - Perona-Malik, 3 - Tukey Biweight + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Perona, P. and Malik, J., 1990. Scale-space and edge detection using anisotropic diffusion. IEEE Transactions on pattern analysis and machine intelligence, 12(7), pp.629-639. + * [2] Black, M.J., Sapiro, G., Marimont, D.H. and Heeger, D., 1998. Robust anisotropic diffusion. IEEE Transactions on image processing, 7(3), pp.421-432. + */ + + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float Diffusion_CPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int penaltytype, int dimX, int dimY, int dimZ); +CCPI_EXPORT float LinearDiff2D(float *Input, float *Output, float lambdaPar, float tau, long dimX, long dimY); +CCPI_EXPORT float NonLinearDiff2D(float *Input, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, long dimX, long dimY); +CCPI_EXPORT float LinearDiff3D(float *Input, float *Output, float lambdaPar, float tau, long dimX, long dimY, long dimZ); +CCPI_EXPORT float NonLinearDiff3D(float *Input, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, long dimX, long dimY, long dimZ); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/FGP_TV_core.c b/src/Core/regularisers_CPU/FGP_TV_core.c new file mode 100644 index 0000000..68d58b7 --- /dev/null +++ b/src/Core/regularisers_CPU/FGP_TV_core.c @@ -0,0 +1,321 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "FGP_TV_core.h" + +/* C-OMP implementation of FGP-TV [1] denoising/regularization model (2D/3D case) + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambdaPar - regularization parameter + * 3. Number of iterations + * 4. eplsilon: tolerance constant + * 5. TV-type: methodTV - 'iso' (0) or 'l1' (1) + * 6. nonneg: 'nonnegativity (0 is OFF by default) + * 7. print information: 0 (off) or 1 (on) + * + * Output: + * [1] Filtered/regularized image + * + * This function is based on the Matlab's code and paper by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + */ + +float TV_FGP_CPU_main(float *Input, float *Output, float lambdaPar, int iterationsNumb, float epsil, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ) +{ + int ll; + long j, DimTotal; + float re, re1; + float tk = 1.0f; + float tkp1=1.0f; + int count = 0; + + if (dimZ <= 1) { + /*2D case */ + float *Output_prev=NULL, *P1=NULL, *P2=NULL, *P1_prev=NULL, *P2_prev=NULL, *R1=NULL, *R2=NULL; + DimTotal = (long)(dimX*dimY); + + Output_prev = calloc(DimTotal, sizeof(float)); + P1 = calloc(DimTotal, sizeof(float)); + P2 = calloc(DimTotal, sizeof(float)); + P1_prev = calloc(DimTotal, sizeof(float)); + P2_prev = calloc(DimTotal, sizeof(float)); + R1 = calloc(DimTotal, sizeof(float)); + R2 = calloc(DimTotal, sizeof(float)); + + /* begin iterations */ + for(ll=0; ll<iterationsNumb; ll++) { + + /* computing the gradient of the objective function */ + Obj_func2D(Input, Output, R1, R2, lambdaPar, (long)(dimX), (long)(dimY)); + + /* apply nonnegativity */ + if (nonneg == 1) for(j=0; j<DimTotal; j++) {if (Output[j] < 0.0f) Output[j] = 0.0f;} + + /*Taking a step towards minus of the gradient*/ + Grad_func2D(P1, P2, Output, R1, R2, lambdaPar, (long)(dimX), (long)(dimY)); + + /* projection step */ + Proj_func2D(P1, P2, methodTV, DimTotal); + + /*updating R and t*/ + tkp1 = (1.0f + sqrt(1.0f + 4.0f*tk*tk))*0.5f; + Rupd_func2D(P1, P1_prev, P2, P2_prev, R1, R2, tkp1, tk, DimTotal); + + /* check early stopping criteria */ + re = 0.0f; re1 = 0.0f; + for(j=0; j<DimTotal; j++) + { + re += pow(Output[j] - Output_prev[j],2); + re1 += pow(Output[j],2); + } + re = sqrt(re)/sqrt(re1); + if (re < epsil) count++; + if (count > 4) break; + + /*storing old values*/ + copyIm(Output, Output_prev, (long)(dimX), (long)(dimY), 1l); + copyIm(P1, P1_prev, (long)(dimX), (long)(dimY), 1l); + copyIm(P2, P2_prev, (long)(dimX), (long)(dimY), 1l); + tk = tkp1; + } + if (printM == 1) printf("FGP-TV iterations stopped at iteration %i \n", ll); + free(Output_prev); free(P1); free(P2); free(P1_prev); free(P2_prev); free(R1); free(R2); + } + else { + /*3D case*/ + float *Output_prev=NULL, *P1=NULL, *P2=NULL, *P3=NULL, *P1_prev=NULL, *P2_prev=NULL, *P3_prev=NULL, *R1=NULL, *R2=NULL, *R3=NULL; + DimTotal = (long)(dimX*dimY*dimZ); + + Output_prev = calloc(DimTotal, sizeof(float)); + P1 = calloc(DimTotal, sizeof(float)); + P2 = calloc(DimTotal, sizeof(float)); + P3 = calloc(DimTotal, sizeof(float)); + P1_prev = calloc(DimTotal, sizeof(float)); + P2_prev = calloc(DimTotal, sizeof(float)); + P3_prev = calloc(DimTotal, sizeof(float)); + R1 = calloc(DimTotal, sizeof(float)); + R2 = calloc(DimTotal, sizeof(float)); + R3 = calloc(DimTotal, sizeof(float)); + + /* begin iterations */ + for(ll=0; ll<iterationsNumb; ll++) { + + /* computing the gradient of the objective function */ + Obj_func3D(Input, Output, R1, R2, R3, lambdaPar, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* apply nonnegativity */ + if (nonneg == 1) for(j=0; j<DimTotal; j++) {if (Output[j] < 0.0f) Output[j] = 0.0f;} + + /*Taking a step towards minus of the gradient*/ + Grad_func3D(P1, P2, P3, Output, R1, R2, R3, lambdaPar, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* projection step */ + Proj_func3D(P1, P2, P3, methodTV, DimTotal); + + /*updating R and t*/ + tkp1 = (1.0f + sqrt(1.0f + 4.0f*tk*tk))*0.5f; + Rupd_func3D(P1, P1_prev, P2, P2_prev, P3, P3_prev, R1, R2, R3, tkp1, tk, DimTotal); + + /* calculate norm - stopping rules*/ + re = 0.0f; re1 = 0.0f; + for(j=0; j<DimTotal; j++) + { + re += pow(Output[j] - Output_prev[j],2); + re1 += pow(Output[j],2); + } + re = sqrt(re)/sqrt(re1); + /* stop if the norm residual is less than the tolerance EPS */ + if (re < epsil) count++; + if (count > 4) break; + + /*storing old values*/ + copyIm(Output, Output_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(P1, P1_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(P2, P2_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(P3, P3_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + tk = tkp1; + } + if (printM == 1) printf("FGP-TV iterations stopped at iteration %i \n", ll); + free(Output_prev); free(P1); free(P2); free(P3); free(P1_prev); free(P2_prev); free(P3_prev); free(R1); free(R2); free(R3); + } + return *Output; +} + +float Obj_func2D(float *A, float *D, float *R1, float *R2, float lambda, long dimX, long dimY) +{ + float val1, val2; + long i,j,index; +#pragma omp parallel for shared(A,D,R1,R2) private(index,i,j,val1,val2) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + /* boundary conditions */ + if (i == 0) {val1 = 0.0f;} else {val1 = R1[j*dimX + (i-1)];} + if (j == 0) {val2 = 0.0f;} else {val2 = R2[(j-1)*dimX + i];} + D[index] = A[index] - lambda*(R1[index] + R2[index] - val1 - val2); + }} + return *D; +} +float Grad_func2D(float *P1, float *P2, float *D, float *R1, float *R2, float lambda, long dimX, long dimY) +{ + float val1, val2, multip; + long i,j,index; + multip = (1.0f/(8.0f*lambda)); +#pragma omp parallel for shared(P1,P2,D,R1,R2,multip) private(index,i,j,val1,val2) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + /* boundary conditions */ + if (i == dimX-1) val1 = 0.0f; else val1 = D[index] - D[j*dimX + (i+1)]; + if (j == dimY-1) val2 = 0.0f; else val2 = D[index] - D[(j+1)*dimX + i]; + P1[index] = R1[index] + multip*val1; + P2[index] = R2[index] + multip*val2; + }} + return 1; +} +float Proj_func2D(float *P1, float *P2, int methTV, long DimTotal) +{ + float val1, val2, denom, sq_denom; + long i; + if (methTV == 0) { + /* isotropic TV*/ +#pragma omp parallel for shared(P1,P2) private(i,denom,sq_denom) + for(i=0; i<DimTotal; i++) { + denom = powf(P1[i],2) + powf(P2[i],2); + if (denom > 1.0f) { + sq_denom = 1.0f/sqrtf(denom); + P1[i] = P1[i]*sq_denom; + P2[i] = P2[i]*sq_denom; + } + } + } + else { + /* anisotropic TV*/ +#pragma omp parallel for shared(P1,P2) private(i,val1,val2) + for(i=0; i<DimTotal; i++) { + val1 = fabs(P1[i]); + val2 = fabs(P2[i]); + if (val1 < 1.0f) {val1 = 1.0f;} + if (val2 < 1.0f) {val2 = 1.0f;} + P1[i] = P1[i]/val1; + P2[i] = P2[i]/val2; + } + } + return 1; +} +float Rupd_func2D(float *P1, float *P1_old, float *P2, float *P2_old, float *R1, float *R2, float tkp1, float tk, long DimTotal) +{ + long i; + float multip; + multip = ((tk-1.0f)/tkp1); +#pragma omp parallel for shared(P1,P2,P1_old,P2_old,R1,R2,multip) private(i) + for(i=0; i<DimTotal; i++) { + R1[i] = P1[i] + multip*(P1[i] - P1_old[i]); + R2[i] = P2[i] + multip*(P2[i] - P2_old[i]); + } + return 1; +} + +/* 3D-case related Functions */ +/*****************************************************************/ +float Obj_func3D(float *A, float *D, float *R1, float *R2, float *R3, float lambda, long dimX, long dimY, long dimZ) +{ + float val1, val2, val3; + long i,j,k,index; +#pragma omp parallel for shared(A,D,R1,R2,R3) private(index,i,j,k,val1,val2,val3) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* boundary conditions */ + if (i == 0) {val1 = 0.0f;} else {val1 = R1[(dimX*dimY)*k + j*dimX + (i-1)];} + if (j == 0) {val2 = 0.0f;} else {val2 = R2[(dimX*dimY)*k + (j-1)*dimX + i];} + if (k == 0) {val3 = 0.0f;} else {val3 = R3[(dimX*dimY)*(k-1) + j*dimX + i];} + D[index] = A[index] - lambda*(R1[index] + R2[index] + R3[index] - val1 - val2 - val3); + }}} + return *D; +} +float Grad_func3D(float *P1, float *P2, float *P3, float *D, float *R1, float *R2, float *R3, float lambda, long dimX, long dimY, long dimZ) +{ + float val1, val2, val3, multip; + long i,j,k, index; + multip = (1.0f/(26.0f*lambda)); +#pragma omp parallel for shared(P1,P2,P3,D,R1,R2,R3,multip) private(index,i,j,k,val1,val2,val3) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* boundary conditions */ + if (i == dimX-1) val1 = 0.0f; else val1 = D[index] - D[(dimX*dimY)*k + j*dimX + (i+1)]; + if (j == dimY-1) val2 = 0.0f; else val2 = D[index] - D[(dimX*dimY)*k + (j+1)*dimX + i]; + if (k == dimZ-1) val3 = 0.0f; else val3 = D[index] - D[(dimX*dimY)*(k+1) + j*dimX + i]; + P1[index] = R1[index] + multip*val1; + P2[index] = R2[index] + multip*val2; + P3[index] = R3[index] + multip*val3; + }}} + return 1; +} +float Proj_func3D(float *P1, float *P2, float *P3, int methTV, long DimTotal) +{ + float val1, val2, val3, denom, sq_denom; + long i; + if (methTV == 0) { + /* isotropic TV*/ + #pragma omp parallel for shared(P1,P2,P3) private(i,val1,val2,val3,sq_denom) + for(i=0; i<DimTotal; i++) { + denom = powf(P1[i],2) + powf(P2[i],2) + powf(P3[i],2); + if (denom > 1.0f) { + sq_denom = 1.0f/sqrtf(denom); + P1[i] = P1[i]*sq_denom; + P2[i] = P2[i]*sq_denom; + P3[i] = P3[i]*sq_denom; + } + } + } + else { + /* anisotropic TV*/ +#pragma omp parallel for shared(P1,P2,P3) private(i,val1,val2,val3) + for(i=0; i<DimTotal; i++) { + val1 = fabs(P1[i]); + val2 = fabs(P2[i]); + val3 = fabs(P3[i]); + if (val1 < 1.0f) {val1 = 1.0f;} + if (val2 < 1.0f) {val2 = 1.0f;} + if (val3 < 1.0f) {val3 = 1.0f;} + P1[i] = P1[i]/val1; + P2[i] = P2[i]/val2; + P3[i] = P3[i]/val3; + } + } + return 1; +} +float Rupd_func3D(float *P1, float *P1_old, float *P2, float *P2_old, float *P3, float *P3_old, float *R1, float *R2, float *R3, float tkp1, float tk, long DimTotal) +{ + long i; + float multip; + multip = ((tk-1.0f)/tkp1); +#pragma omp parallel for shared(P1,P2,P3,P1_old,P2_old,P3_old,R1,R2,R3,multip) private(i) + for(i=0; i<DimTotal; i++) { + R1[i] = P1[i] + multip*(P1[i] - P1_old[i]); + R2[i] = P2[i] + multip*(P2[i] - P2_old[i]); + R3[i] = P3[i] + multip*(P3[i] - P3_old[i]); + } + return 1; +} diff --git a/src/Core/regularisers_CPU/FGP_TV_core.h b/src/Core/regularisers_CPU/FGP_TV_core.h new file mode 100644 index 0000000..3418604 --- /dev/null +++ b/src/Core/regularisers_CPU/FGP_TV_core.h @@ -0,0 +1,63 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +//#include <matrix.h> +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + +/* C-OMP implementation of FGP-TV [1] denoising/regularization model (2D/3D case) + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambda - regularization parameter + * 3. Number of iterations + * 4. eplsilon: tolerance constant + * 5. TV-type: methodTV - 'iso' (0) or 'l1' (1) + * 6. nonneg: 'nonnegativity (0 is OFF by default) + * 7. print information: 0 (off) or 1 (on) + * + * Output: + * [1] Filtered/regularized image + * + * This function is based on the Matlab's code and paper by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + */ + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float TV_FGP_CPU_main(float *Input, float *Output, float lambdaPar, int iterationsNumb, float epsil, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ); + +CCPI_EXPORT float Obj_func2D(float *A, float *D, float *R1, float *R2, float lambda, long dimX, long dimY); +CCPI_EXPORT float Grad_func2D(float *P1, float *P2, float *D, float *R1, float *R2, float lambda, long dimX, long dimY); +CCPI_EXPORT float Proj_func2D(float *P1, float *P2, int methTV, long DimTotal); +CCPI_EXPORT float Rupd_func2D(float *P1, float *P1_old, float *P2, float *P2_old, float *R1, float *R2, float tkp1, float tk, long DimTotal); + +CCPI_EXPORT float Obj_func3D(float *A, float *D, float *R1, float *R2, float *R3, float lambda, long dimX, long dimY, long dimZ); +CCPI_EXPORT float Grad_func3D(float *P1, float *P2, float *P3, float *D, float *R1, float *R2, float *R3, float lambda, long dimX, long dimY, long dimZ); +CCPI_EXPORT float Proj_func3D(float *P1, float *P2, float *P3, int methTV, long DimTotal); +CCPI_EXPORT float Rupd_func3D(float *P1, float *P1_old, float *P2, float *P2_old, float *P3, float *P3_old, float *R1, float *R2, float *R3, float tkp1, float tk, long DimTotal); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/FGP_dTV_core.c b/src/Core/regularisers_CPU/FGP_dTV_core.c new file mode 100644 index 0000000..17b75ff --- /dev/null +++ b/src/Core/regularisers_CPU/FGP_dTV_core.c @@ -0,0 +1,441 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "FGP_dTV_core.h" + +/* C-OMP implementation of FGP-dTV [1,2] denoising/regularization model (2D/3D case) + * which employs structural similarity of the level sets of two images/volumes, see [1,2] + * The current implementation updates image 1 while image 2 is being fixed. + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. Additional reference image/volume of the same dimensions as (1) [REQUIRED] + * 3. lambdaPar - regularization parameter [REQUIRED] + * 4. Number of iterations [OPTIONAL] + * 5. eplsilon: tolerance constant [OPTIONAL] + * 6. eta: smoothing constant to calculate gradient of the reference [OPTIONAL] * + * 7. TV-type: methodTV - 'iso' (0) or 'l1' (1) [OPTIONAL] + * 8. nonneg: 'nonnegativity (0 is OFF by default) [OPTIONAL] + * 9. print information: 0 (off) or 1 (on) [OPTIONAL] + * + * Output: + * [1] Filtered/regularized image/volume + * + * This function is based on the Matlab's codes and papers by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + * [2] M. J. Ehrhardt and M. M. Betcke, Multi-Contrast MRI Reconstruction with Structure-Guided Total Variation, SIAM Journal on Imaging Sciences 9(3), pp. 1084–1106 + */ + +float dTV_FGP_CPU_main(float *Input, float *InputRef, float *Output, float lambdaPar, int iterationsNumb, float epsil, float eta, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ) +{ + int ll; + long j, DimTotal; + float re, re1; + float tk = 1.0f; + float tkp1=1.0f; + int count = 0; + + if (dimZ <= 1) { + /*2D case */ + float *Output_prev=NULL, *P1=NULL, *P2=NULL, *P1_prev=NULL, *P2_prev=NULL, *R1=NULL, *R2=NULL, *InputRef_x=NULL, *InputRef_y=NULL; + DimTotal = (long)(dimX*dimY); + + Output_prev = calloc(DimTotal, sizeof(float)); + P1 = calloc(DimTotal, sizeof(float)); + P2 = calloc(DimTotal, sizeof(float)); + P1_prev = calloc(DimTotal, sizeof(float)); + P2_prev = calloc(DimTotal, sizeof(float)); + R1 = calloc(DimTotal, sizeof(float)); + R2 = calloc(DimTotal, sizeof(float)); + InputRef_x = calloc(DimTotal, sizeof(float)); + InputRef_y = calloc(DimTotal, sizeof(float)); + + /* calculate gradient field (smoothed) for the reference image */ + GradNorm_func2D(InputRef, InputRef_x, InputRef_y, eta, (long)(dimX), (long)(dimY)); + + /* begin iterations */ + for(ll=0; ll<iterationsNumb; ll++) { + + /*projects a 2D vector field R-1,2 onto the orthogonal complement of another 2D vector field InputRef_xy*/ + ProjectVect_func2D(R1, R2, InputRef_x, InputRef_y, (long)(dimX), (long)(dimY)); + + /* computing the gradient of the objective function */ + Obj_dfunc2D(Input, Output, R1, R2, lambdaPar, (long)(dimX), (long)(dimY)); + + /* apply nonnegativity */ + if (nonneg == 1) for(j=0; j<DimTotal; j++) {if (Output[j] < 0.0f) Output[j] = 0.0f;} + + /*Taking a step towards minus of the gradient*/ + Grad_dfunc2D(P1, P2, Output, R1, R2, InputRef_x, InputRef_y, lambdaPar, (long)(dimX), (long)(dimY)); + + /* projection step */ + Proj_dfunc2D(P1, P2, methodTV, DimTotal); + + /*updating R and t*/ + tkp1 = (1.0f + sqrt(1.0f + 4.0f*tk*tk))*0.5f; + Rupd_dfunc2D(P1, P1_prev, P2, P2_prev, R1, R2, tkp1, tk, DimTotal); + + /* check early stopping criteria */ + re = 0.0f; re1 = 0.0f; + for(j=0; j<DimTotal; j++) + { + re += pow(Output[j] - Output_prev[j],2); + re1 += pow(Output[j],2); + } + re = sqrt(re)/sqrt(re1); + if (re < epsil) count++; + if (count > 4) break; + + /*storing old values*/ + copyIm(Output, Output_prev, (long)(dimX), (long)(dimY), 1l); + copyIm(P1, P1_prev, (long)(dimX), (long)(dimY), 1l); + copyIm(P2, P2_prev, (long)(dimX), (long)(dimY), 1l); + tk = tkp1; + } + if (printM == 1) printf("FGP-dTV iterations stopped at iteration %i \n", ll); + free(Output_prev); free(P1); free(P2); free(P1_prev); free(P2_prev); free(R1); free(R2); free(InputRef_x); free(InputRef_y); + } + else { + /*3D case*/ + float *Output_prev=NULL, *P1=NULL, *P2=NULL, *P3=NULL, *P1_prev=NULL, *P2_prev=NULL, *P3_prev=NULL, *R1=NULL, *R2=NULL, *R3=NULL, *InputRef_x=NULL, *InputRef_y=NULL, *InputRef_z=NULL; + DimTotal = (long)(dimX*dimY*dimZ); + + Output_prev = calloc(DimTotal, sizeof(float)); + P1 = calloc(DimTotal, sizeof(float)); + P2 = calloc(DimTotal, sizeof(float)); + P3 = calloc(DimTotal, sizeof(float)); + P1_prev = calloc(DimTotal, sizeof(float)); + P2_prev = calloc(DimTotal, sizeof(float)); + P3_prev = calloc(DimTotal, sizeof(float)); + R1 = calloc(DimTotal, sizeof(float)); + R2 = calloc(DimTotal, sizeof(float)); + R3 = calloc(DimTotal, sizeof(float)); + InputRef_x = calloc(DimTotal, sizeof(float)); + InputRef_y = calloc(DimTotal, sizeof(float)); + InputRef_z = calloc(DimTotal, sizeof(float)); + + /* calculate gradient field (smoothed) for the reference volume */ + GradNorm_func3D(InputRef, InputRef_x, InputRef_y, InputRef_z, eta, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* begin iterations */ + for(ll=0; ll<iterationsNumb; ll++) { + + /*projects a 3D vector field R-1,2,3 onto the orthogonal complement of another 3D vector field InputRef_xyz*/ + ProjectVect_func3D(R1, R2, R3, InputRef_x, InputRef_y, InputRef_z, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* computing the gradient of the objective function */ + Obj_dfunc3D(Input, Output, R1, R2, R3, lambdaPar, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* apply nonnegativity */ + if (nonneg == 1) for(j=0; j<DimTotal; j++) {if (Output[j] < 0.0f) Output[j] = 0.0f;} + + /*Taking a step towards minus of the gradient*/ + Grad_dfunc3D(P1, P2, P3, Output, R1, R2, R3, InputRef_x, InputRef_y, InputRef_z, lambdaPar, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* projection step */ + Proj_dfunc3D(P1, P2, P3, methodTV, DimTotal); + + /*updating R and t*/ + tkp1 = (1.0f + sqrt(1.0f + 4.0f*tk*tk))*0.5f; + Rupd_dfunc3D(P1, P1_prev, P2, P2_prev, P3, P3_prev, R1, R2, R3, tkp1, tk, DimTotal); + + /* calculate norm - stopping rules*/ + re = 0.0f; re1 = 0.0f; + for(j=0; j<DimTotal; j++) + { + re += pow(Output[j] - Output_prev[j],2); + re1 += pow(Output[j],2); + } + re = sqrt(re)/sqrt(re1); + /* stop if the norm residual is less than the tolerance EPS */ + if (re < epsil) count++; + if (count > 4) break; + + /*storing old values*/ + copyIm(Output, Output_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(P1, P1_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(P2, P2_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(P3, P3_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + tk = tkp1; + } + if (printM == 1) printf("FGP-dTV iterations stopped at iteration %i \n", ll); + free(Output_prev); free(P1); free(P2); free(P3); free(P1_prev); free(P2_prev); free(P3_prev); free(R1); free(R2); free(R3); free(InputRef_x); free(InputRef_y); free(InputRef_z); + } + return *Output; +} + + +/********************************************************************/ +/***************************2D Functions*****************************/ +/********************************************************************/ + +float GradNorm_func2D(float *B, float *B_x, float *B_y, float eta, long dimX, long dimY) +{ + long i,j,index; + float val1, val2, gradX, gradY, magn; +#pragma omp parallel for shared(B, B_x, B_y) private(i,j,index,val1,val2,gradX,gradY,magn) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + /* zero boundary conditions */ + if (i == dimX-1) {val1 = 0.0f;} else {val1 = B[j*dimX + (i+1)];} + if (j == dimY-1) {val2 = 0.0f;} else {val2 = B[(j+1)*dimX + i];} + gradX = val1 - B[index]; + gradY = val2 - B[index]; + magn = pow(gradX,2) + pow(gradY,2); + magn = sqrt(magn + pow(eta,2)); /* the eta-smoothed gradients magnitude */ + B_x[index] = gradX/magn; + B_y[index] = gradY/magn; + }} + return 1; +} + +float ProjectVect_func2D(float *R1, float *R2, float *B_x, float *B_y, long dimX, long dimY) +{ + long i,j,index; + float in_prod; +#pragma omp parallel for shared(R1, R2, B_x, B_y) private(index,i,j,in_prod) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + in_prod = R1[index]*B_x[index] + R2[index]*B_y[index]; /* calculate inner product */ + R1[index] = R1[index] - in_prod*B_x[index]; + R2[index] = R2[index] - in_prod*B_y[index]; + }} + return 1; +} + +float Obj_dfunc2D(float *A, float *D, float *R1, float *R2, float lambda, long dimX, long dimY) +{ + float val1, val2; + long i,j,index; +#pragma omp parallel for shared(A,D,R1,R2) private(index,i,j,val1,val2) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + /* boundary conditions */ + if (i == 0) {val1 = 0.0f;} else {val1 = R1[j*dimX + (i-1)];} + if (j == 0) {val2 = 0.0f;} else {val2 = R2[(j-1)*dimX + i];} + D[index] = A[index] - lambda*(R1[index] + R2[index] - val1 - val2); + }} + return *D; +} +float Grad_dfunc2D(float *P1, float *P2, float *D, float *R1, float *R2, float *B_x, float *B_y, float lambda, long dimX, long dimY) +{ + float val1, val2, multip, in_prod; + long i,j,index; + multip = (1.0f/(8.0f*lambda)); +#pragma omp parallel for shared(P1,P2,D,R1,R2,B_x,B_y,multip) private(i,j,index,val1,val2,in_prod) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + /* boundary conditions */ + if (i == dimX-1) val1 = 0.0f; else val1 = D[index] - D[j*dimX + (i+1)]; + if (j == dimY-1) val2 = 0.0f; else val2 = D[index] - D[(j+1)*dimX + i]; + + in_prod = val1*B_x[index] + val2*B_y[index]; /* calculate inner product */ + val1 = val1 - in_prod*B_x[index]; + val2 = val2 - in_prod*B_y[index]; + + P1[index] = R1[index] + multip*val1; + P2[index] = R2[index] + multip*val2; + + }} + return 1; +} +float Proj_dfunc2D(float *P1, float *P2, int methTV, long DimTotal) +{ + float val1, val2, denom, sq_denom; + long i; + if (methTV == 0) { + /* isotropic TV*/ +#pragma omp parallel for shared(P1,P2) private(i,denom,sq_denom) + for(i=0; i<DimTotal; i++) { + denom = powf(P1[i],2) + powf(P2[i],2); + if (denom > 1.0f) { + sq_denom = 1.0f/sqrtf(denom); + P1[i] = P1[i]*sq_denom; + P2[i] = P2[i]*sq_denom; + } + } + } + else { + /* anisotropic TV*/ +#pragma omp parallel for shared(P1,P2) private(i,val1,val2) + for(i=0; i<DimTotal; i++) { + val1 = fabs(P1[i]); + val2 = fabs(P2[i]); + if (val1 < 1.0f) {val1 = 1.0f;} + if (val2 < 1.0f) {val2 = 1.0f;} + P1[i] = P1[i]/val1; + P2[i] = P2[i]/val2; + } + } + return 1; +} +float Rupd_dfunc2D(float *P1, float *P1_old, float *P2, float *P2_old, float *R1, float *R2, float tkp1, float tk, long DimTotal) +{ + long i; + float multip; + multip = ((tk-1.0f)/tkp1); +#pragma omp parallel for shared(P1,P2,P1_old,P2_old,R1,R2,multip) private(i) + for(i=0; i<DimTotal; i++) { + R1[i] = P1[i] + multip*(P1[i] - P1_old[i]); + R2[i] = P2[i] + multip*(P2[i] - P2_old[i]); + } + return 1; +} + +/********************************************************************/ +/***************************3D Functions*****************************/ +/********************************************************************/ +float GradNorm_func3D(float *B, float *B_x, float *B_y, float *B_z, float eta, long dimX, long dimY, long dimZ) +{ + long i, j, k, index; + float val1, val2, val3, gradX, gradY, gradZ, magn; +#pragma omp parallel for shared(B, B_x, B_y, B_z) private(i,j,k,index,val1,val2,val3,gradX,gradY,gradZ,magn) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + + /* zero boundary conditions */ + if (i == dimX-1) {val1 = 0.0f;} else {val1 = B[(dimX*dimY)*k + j*dimX+(i+1)];} + if (j == dimY-1) {val2 = 0.0f;} else {val2 = B[(dimX*dimY)*k + (j+1)*dimX+i];} + if (k == dimZ-1) {val3 = 0.0f;} else {val3 = B[(dimX*dimY)*(k+1) + (j)*dimX+i];} + + gradX = val1 - B[index]; + gradY = val2 - B[index]; + gradZ = val3 - B[index]; + magn = pow(gradX,2) + pow(gradY,2) + pow(gradZ,2); + magn = sqrt(magn + pow(eta,2)); /* the eta-smoothed gradients magnitude */ + B_x[index] = gradX/magn; + B_y[index] = gradY/magn; + B_z[index] = gradZ/magn; + }}} + return 1; +} + +float ProjectVect_func3D(float *R1, float *R2, float *R3, float *B_x, float *B_y, float *B_z, long dimX, long dimY, long dimZ) +{ + long i,j,k,index; + float in_prod; +#pragma omp parallel for shared(R1, R2, R3, B_x, B_y, B_z) private(index,i,j,k,in_prod) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + in_prod = R1[index]*B_x[index] + R2[index]*B_y[index] + R3[index]*B_z[index]; /* calculate inner product */ + R1[index] = R1[index] - in_prod*B_x[index]; + R2[index] = R2[index] - in_prod*B_y[index]; + R3[index] = R3[index] - in_prod*B_z[index]; + }}} + return 1; +} + +float Obj_dfunc3D(float *A, float *D, float *R1, float *R2, float *R3, float lambda, long dimX, long dimY, long dimZ) +{ + float val1, val2, val3; + long i,j,k,index; +#pragma omp parallel for shared(A,D,R1,R2,R3) private(index,i,j,k,val1,val2,val3) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* boundary conditions */ + if (i == 0) {val1 = 0.0f;} else {val1 = R1[(dimX*dimY)*k + j*dimX + (i-1)];} + if (j == 0) {val2 = 0.0f;} else {val2 = R2[(dimX*dimY)*k + (j-1)*dimX + i];} + if (k == 0) {val3 = 0.0f;} else {val3 = R3[(dimX*dimY)*(k-1) + j*dimX + i];} + D[index] = A[index] - lambda*(R1[index] + R2[index] + R3[index] - val1 - val2 - val3); + }}} + return *D; +} +float Grad_dfunc3D(float *P1, float *P2, float *P3, float *D, float *R1, float *R2, float *R3, float *B_x, float *B_y, float *B_z, float lambda, long dimX, long dimY, long dimZ) +{ + float val1, val2, val3, multip, in_prod; + long i,j,k, index; + multip = (1.0f/(26.0f*lambda)); +#pragma omp parallel for shared(P1,P2,P3,D,R1,R2,R3,multip) private(index,i,j,k,val1,val2,val3,in_prod) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* boundary conditions */ + if (i == dimX-1) val1 = 0.0f; else val1 = D[index] - D[(dimX*dimY)*k + j*dimX + (i+1)]; + if (j == dimY-1) val2 = 0.0f; else val2 = D[index] - D[(dimX*dimY)*k + (j+1)*dimX + i]; + if (k == dimZ-1) val3 = 0.0f; else val3 = D[index] - D[(dimX*dimY)*(k+1) + j*dimX + i]; + + in_prod = val1*B_x[index] + val2*B_y[index] + val3*B_z[index]; /* calculate inner product */ + val1 = val1 - in_prod*B_x[index]; + val2 = val2 - in_prod*B_y[index]; + val3 = val3 - in_prod*B_z[index]; + + P1[index] = R1[index] + multip*val1; + P2[index] = R2[index] + multip*val2; + P3[index] = R3[index] + multip*val3; + }}} + return 1; +} +float Proj_dfunc3D(float *P1, float *P2, float *P3, int methTV, long DimTotal) +{ + float val1, val2, val3, denom, sq_denom; + long i; + if (methTV == 0) { + /* isotropic TV*/ + #pragma omp parallel for shared(P1,P2,P3) private(i,val1,val2,val3,sq_denom) + for(i=0; i<DimTotal; i++) { + denom = powf(P1[i],2) + powf(P2[i],2) + powf(P3[i],2); + if (denom > 1.0f) { + sq_denom = 1.0f/sqrtf(denom); + P1[i] = P1[i]*sq_denom; + P2[i] = P2[i]*sq_denom; + P3[i] = P3[i]*sq_denom; + } + } + } + else { + /* anisotropic TV*/ +#pragma omp parallel for shared(P1,P2,P3) private(i,val1,val2,val3) + for(i=0; i<DimTotal; i++) { + val1 = fabs(P1[i]); + val2 = fabs(P2[i]); + val3 = fabs(P3[i]); + if (val1 < 1.0f) {val1 = 1.0f;} + if (val2 < 1.0f) {val2 = 1.0f;} + if (val3 < 1.0f) {val3 = 1.0f;} + P1[i] = P1[i]/val1; + P2[i] = P2[i]/val2; + P3[i] = P3[i]/val3; + } + } + return 1; +} +float Rupd_dfunc3D(float *P1, float *P1_old, float *P2, float *P2_old, float *P3, float *P3_old, float *R1, float *R2, float *R3, float tkp1, float tk, long DimTotal) +{ + long i; + float multip; + multip = ((tk-1.0f)/tkp1); +#pragma omp parallel for shared(P1,P2,P3,P1_old,P2_old,P3_old,R1,R2,R3,multip) private(i) + for(i=0; i<DimTotal; i++) { + R1[i] = P1[i] + multip*(P1[i] - P1_old[i]); + R2[i] = P2[i] + multip*(P2[i] - P2_old[i]); + R3[i] = P3[i] + multip*(P3[i] - P3_old[i]); + } + return 1; +} diff --git a/src/Core/regularisers_CPU/FGP_dTV_core.h b/src/Core/regularisers_CPU/FGP_dTV_core.h new file mode 100644 index 0000000..442dd30 --- /dev/null +++ b/src/Core/regularisers_CPU/FGP_dTV_core.h @@ -0,0 +1,72 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +//#include <matrix.h> +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + +/* C-OMP implementation of FGP-dTV [1,2] denoising/regularization model (2D/3D case) + * which employs structural similarity of the level sets of two images/volumes, see [1,2] + * The current implementation updates image 1 while image 2 is being fixed. + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. Additional reference image/volume of the same dimensions as (1) [REQUIRED] + * 3. lambdaPar - regularization parameter [REQUIRED] + * 4. Number of iterations [OPTIONAL] + * 5. eplsilon: tolerance constant [OPTIONAL] + * 6. eta: smoothing constant to calculate gradient of the reference [OPTIONAL] * + * 7. TV-type: methodTV - 'iso' (0) or 'l1' (1) [OPTIONAL] + * 8. nonneg: 'nonnegativity (0 is OFF by default) [OPTIONAL] + * 9. print information: 0 (off) or 1 (on) [OPTIONAL] + * + * Output: + * [1] Filtered/regularized image/volume + * + * This function is based on the Matlab's codes and papers by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + * [2] M. J. Ehrhardt and M. M. Betcke, Multi-Contrast MRI Reconstruction with Structure-Guided Total Variation, SIAM Journal on Imaging Sciences 9(3), pp. 1084–1106 + */ + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float dTV_FGP_CPU_main(float *Input, float *InputRef, float *Output, float lambdaPar, int iterationsNumb, float epsil, float eta, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ); + +CCPI_EXPORT float GradNorm_func2D(float *B, float *B_x, float *B_y, float eta, long dimX, long dimY); +CCPI_EXPORT float ProjectVect_func2D(float *R1, float *R2, float *B_x, float *B_y, long dimX, long dimY); +CCPI_EXPORT float Obj_dfunc2D(float *A, float *D, float *R1, float *R2, float lambda, long dimX, long dimY); +CCPI_EXPORT float Grad_dfunc2D(float *P1, float *P2, float *D, float *R1, float *R2, float *B_x, float *B_y, float lambda, long dimX, long dimY); +CCPI_EXPORT float Proj_dfunc2D(float *P1, float *P2, int methTV, long DimTotal); +CCPI_EXPORT float Rupd_dfunc2D(float *P1, float *P1_old, float *P2, float *P2_old, float *R1, float *R2, float tkp1, float tk, long DimTotal); + +CCPI_EXPORT float GradNorm_func3D(float *B, float *B_x, float *B_y, float *B_z, float eta, long dimX, long dimY, long dimZ); +CCPI_EXPORT float ProjectVect_func3D(float *R1, float *R2, float *R3, float *B_x, float *B_y, float *B_z, long dimX, long dimY, long dimZ); +CCPI_EXPORT float Obj_dfunc3D(float *A, float *D, float *R1, float *R2, float *R3, float lambda, long dimX, long dimY, long dimZ); +CCPI_EXPORT float Grad_dfunc3D(float *P1, float *P2, float *P3, float *D, float *R1, float *R2, float *R3, float *B_x, float *B_y, float *B_z, float lambda, long dimX, long dimY, long dimZ); +CCPI_EXPORT float Proj_dfunc3D(float *P1, float *P2, float *P3, int methTV, long DimTotal); +CCPI_EXPORT float Rupd_dfunc3D(float *P1, float *P1_old, float *P2, float *P2_old, float *P3, float *P3_old, float *R1, float *R2, float *R3, float tkp1, float tk, long DimTotal); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/LLT_ROF_core.c b/src/Core/regularisers_CPU/LLT_ROF_core.c new file mode 100644 index 0000000..8416a14 --- /dev/null +++ b/src/Core/regularisers_CPU/LLT_ROF_core.c @@ -0,0 +1,410 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "LLT_ROF_core.h" +#define EPS_LLT 0.01 +#define EPS_ROF 1.0e-12 +#define MAX(x, y) (((x) > (y)) ? (x) : (y)) +#define MIN(x, y) (((x) < (y)) ? (x) : (y)) + +/*sign function*/ +int signLLT(float x) { + return (x > 0) - (x < 0); +} + +/* C-OMP implementation of Lysaker, Lundervold and Tai (LLT) model [1] combined with Rudin-Osher-Fatemi [2] TV regularisation penalty. + * +* This penalty can deliver visually pleasant piecewise-smooth recovery if regularisation parameters are selected well. +* The rule of thumb for selection is to start with lambdaLLT = 0 (just the ROF-TV model) and then proceed to increase +* lambdaLLT starting with smaller values. +* +* Input Parameters: +* 1. U0 - original noise image/volume +* 2. lambdaROF - ROF-related regularisation parameter +* 3. lambdaLLT - LLT-related regularisation parameter +* 4. tau - time-marching step +* 5. iter - iterations number (for both models) +* +* Output: +* Filtered/regularised image +* +* References: +* [1] Lysaker, M., Lundervold, A. and Tai, X.C., 2003. Noise removal using fourth-order partial differential equation with applications to medical magnetic resonance images in space and time. IEEE Transactions on image processing, 12(12), pp.1579-1590. +* [2] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" +*/ + +float LLT_ROF_CPU_main(float *Input, float *Output, float lambdaROF, float lambdaLLT, int iterationsNumb, float tau, int dimX, int dimY, int dimZ) +{ + long DimTotal; + int ll; + float *D1_LLT=NULL, *D2_LLT=NULL, *D3_LLT=NULL, *D1_ROF=NULL, *D2_ROF=NULL, *D3_ROF=NULL; + + DimTotal = (long)(dimX*dimY*dimZ); + + D1_ROF = calloc(DimTotal, sizeof(float)); + D2_ROF = calloc(DimTotal, sizeof(float)); + D3_ROF = calloc(DimTotal, sizeof(float)); + + D1_LLT = calloc(DimTotal, sizeof(float)); + D2_LLT = calloc(DimTotal, sizeof(float)); + D3_LLT = calloc(DimTotal, sizeof(float)); + + copyIm(Input, Output, (long)(dimX), (long)(dimY), (long)(dimZ)); /* initialize */ + + for(ll = 0; ll < iterationsNumb; ll++) { + if (dimZ == 1) { + /* 2D case */ + /****************ROF******************/ + /* calculate first-order differences */ + D1_func_ROF(Output, D1_ROF, (long)(dimX), (long)(dimY), 1l); + D2_func_ROF(Output, D2_ROF, (long)(dimX), (long)(dimY), 1l); + /****************LLT******************/ + /* estimate second-order derrivatives */ + der2D_LLT(Output, D1_LLT, D2_LLT, (long)(dimX), (long)(dimY), 1l); + /* Joint update for ROF and LLT models */ + Update2D_LLT_ROF(Input, Output, D1_LLT, D2_LLT, D1_ROF, D2_ROF, lambdaROF, lambdaLLT, tau, (long)(dimX), (long)(dimY), 1l); + } + else { + /* 3D case */ + /* calculate first-order differences */ + D1_func_ROF(Output, D1_ROF, (long)(dimX), (long)(dimY), (long)(dimZ)); + D2_func_ROF(Output, D2_ROF, (long)(dimX), (long)(dimY), (long)(dimZ)); + D3_func_ROF(Output, D3_ROF, (long)(dimX), (long)(dimY), (long)(dimZ)); + /****************LLT******************/ + /* estimate second-order derrivatives */ + der3D_LLT(Output, D1_LLT, D2_LLT, D3_LLT,(long)(dimX), (long)(dimY), (long)(dimZ)); + /* Joint update for ROF and LLT models */ + Update3D_LLT_ROF(Input, Output, D1_LLT, D2_LLT, D3_LLT, D1_ROF, D2_ROF, D3_ROF, lambdaROF, lambdaLLT, tau, (long)(dimX), (long)(dimY), (long)(dimZ)); + } + } /*end of iterations*/ + free(D1_LLT);free(D2_LLT);free(D3_LLT); + free(D1_ROF);free(D2_ROF);free(D3_ROF); + return *Output; +} + +/*************************************************************************/ +/**********************LLT-related functions *****************************/ +/*************************************************************************/ +float der2D_LLT(float *U, float *D1, float *D2, long dimX, long dimY, long dimZ) +{ + long i, j, index, i_p, i_m, j_m, j_p; + float dxx, dyy, denom_xx, denom_yy; +#pragma omp parallel for shared(U,D1,D2) private(i, j, index, i_p, i_m, j_m, j_p, denom_xx, denom_yy, dxx, dyy) + for (i = 0; i<dimX; i++) { + for (j = 0; j<dimY; j++) { + index = j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i_p = i + 1; if (i_p == dimX) i_p = i - 1; + i_m = i - 1; if (i_m < 0) i_m = i + 1; + j_p = j + 1; if (j_p == dimY) j_p = j - 1; + j_m = j - 1; if (j_m < 0) j_m = j + 1; + + dxx = U[j*dimX+i_p] - 2.0f*U[index] + U[j*dimX+i_m]; + dyy = U[j_p*dimX+i] - 2.0f*U[index] + U[j_m*dimX+i]; + + denom_xx = fabs(dxx) + EPS_LLT; + denom_yy = fabs(dyy) + EPS_LLT; + + D1[index] = dxx / denom_xx; + D2[index] = dyy / denom_yy; + } + } + return 1; +} + +float der3D_LLT(float *U, float *D1, float *D2, float *D3, long dimX, long dimY, long dimZ) + { + long i, j, k, i_p, i_m, j_m, j_p, k_p, k_m, index; + float dxx, dyy, dzz, denom_xx, denom_yy, denom_zz; + #pragma omp parallel for shared(U,D1,D2,D3) private(i, j, index, k, i_p, i_m, j_m, j_p, k_p, k_m, denom_xx, denom_yy, denom_zz, dxx, dyy, dzz) + for (i = 0; i<dimX; i++) { + for (j = 0; j<dimY; j++) { + for (k = 0; k<dimZ; k++) { + /* symmetric boundary conditions (Neuman) */ + i_p = i + 1; if (i_p == dimX) i_p = i - 1; + i_m = i - 1; if (i_m < 0) i_m = i + 1; + j_p = j + 1; if (j_p == dimY) j_p = j - 1; + j_m = j - 1; if (j_m < 0) j_m = j + 1; + k_p = k + 1; if (k_p == dimZ) k_p = k - 1; + k_m = k - 1; if (k_m < 0) k_m = k + 1; + + index = (dimX*dimY)*k + j*dimX+i; + + dxx = U[(dimX*dimY)*k + j*dimX+i_p] - 2.0f*U[index] + U[(dimX*dimY)*k + j*dimX+i_m]; + dyy = U[(dimX*dimY)*k + j_p*dimX+i] - 2.0f*U[index] + U[(dimX*dimY)*k + j_m*dimX+i]; + dzz = U[(dimX*dimY)*k_p + j*dimX+i] - 2.0f*U[index] + U[(dimX*dimY)*k_m + j*dimX+i]; + + denom_xx = fabs(dxx) + EPS_LLT; + denom_yy = fabs(dyy) + EPS_LLT; + denom_zz = fabs(dzz) + EPS_LLT; + + D1[index] = dxx / denom_xx; + D2[index] = dyy / denom_yy; + D3[index] = dzz / denom_zz; + } + } + } + return 1; + } + +/*************************************************************************/ +/**********************ROF-related functions *****************************/ +/*************************************************************************/ + +/* calculate differences 1 */ +float D1_func_ROF(float *A, float *D1, long dimX, long dimY, long dimZ) +{ + float NOMx_1, NOMy_1, NOMy_0, NOMz_1, NOMz_0, denom1, denom2,denom3, T1; + long i,j,k,i1,i2,k1,j1,j2,k2,index; + + if (dimZ > 1) { +#pragma omp parallel for shared (A, D1, dimX, dimY, dimZ) private(index, i, j, k, i1, j1, k1, i2, j2, k2, NOMx_1,NOMy_1,NOMy_0,NOMz_1,NOMz_0,denom1,denom2,denom3,T1) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /* Forward-backward differences */ + NOMx_1 = A[(dimX*dimY)*k + j1*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[(dimX*dimY)*k + j*dimX + i1] - A[index]; /* y+ */ + /*NOMx_0 = (A[(i)*dimY + j] - A[(i2)*dimY + j]); */ /* x- */ + NOMy_0 = A[index] - A[(dimX*dimY)*k + j*dimX + i2]; /* y- */ + + NOMz_1 = A[(dimX*dimY)*k1 + j*dimX + i] - A[index]; /* z+ */ + NOMz_0 = A[index] - A[(dimX*dimY)*k2 + j*dimX + i]; /* z- */ + + + denom1 = NOMx_1*NOMx_1; + denom2 = 0.5f*(signLLT(NOMy_1) + signLLT(NOMy_0))*(MIN(fabs(NOMy_1),fabs(NOMy_0))); + denom2 = denom2*denom2; + denom3 = 0.5f*(signLLT(NOMz_1) + signLLT(NOMz_0))*(MIN(fabs(NOMz_1),fabs(NOMz_0))); + denom3 = denom3*denom3; + T1 = sqrt(denom1 + denom2 + denom3 + EPS_ROF); + D1[index] = NOMx_1/T1; + }}} + } + else { +#pragma omp parallel for shared (A, D1, dimX, dimY) private(i, j, i1, j1, i2, j2,NOMx_1,NOMy_1,NOMy_0,denom1,denom2,T1,index) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + index = j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + + /* Forward-backward differences */ + NOMx_1 = A[j1*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[j*dimX + i1] - A[index]; /* y+ */ + /*NOMx_0 = (A[(i)*dimY + j] - A[(i2)*dimY + j]); */ /* x- */ + NOMy_0 = A[index] - A[(j)*dimX + i2]; /* y- */ + + denom1 = NOMx_1*NOMx_1; + denom2 = 0.5f*(signLLT(NOMy_1) + signLLT(NOMy_0))*(MIN(fabs(NOMy_1),fabs(NOMy_0))); + denom2 = denom2*denom2; + T1 = sqrtf(denom1 + denom2 + EPS_ROF); + D1[index] = NOMx_1/T1; + }} + } + return *D1; +} +/* calculate differences 2 */ +float D2_func_ROF(float *A, float *D2, long dimX, long dimY, long dimZ) +{ + float NOMx_1, NOMy_1, NOMx_0, NOMz_1, NOMz_0, denom1, denom2, denom3, T2; + long i,j,k,i1,i2,k1,j1,j2,k2,index; + + if (dimZ > 1) { +#pragma omp parallel for shared (A, D2, dimX, dimY, dimZ) private(index, i, j, k, i1, j1, k1, i2, j2, k2, NOMx_1, NOMy_1, NOMx_0, NOMz_1, NOMz_0, denom1, denom2, denom3, T2) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + + /* Forward-backward differences */ + NOMx_1 = A[(dimX*dimY)*k + (j1)*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[(dimX*dimY)*k + (j)*dimX + i1] - A[index]; /* y+ */ + NOMx_0 = A[index] - A[(dimX*dimY)*k + (j2)*dimX + i]; /* x- */ + NOMz_1 = A[(dimX*dimY)*k1 + j*dimX + i] - A[index]; /* z+ */ + NOMz_0 = A[index] - A[(dimX*dimY)*k2 + (j)*dimX + i]; /* z- */ + + + denom1 = NOMy_1*NOMy_1; + denom2 = 0.5f*(signLLT(NOMx_1) + signLLT(NOMx_0))*(MIN(fabs(NOMx_1),fabs(NOMx_0))); + denom2 = denom2*denom2; + denom3 = 0.5f*(signLLT(NOMz_1) + signLLT(NOMz_0))*(MIN(fabs(NOMz_1),fabs(NOMz_0))); + denom3 = denom3*denom3; + T2 = sqrtf(denom1 + denom2 + denom3 + EPS_ROF); + D2[index] = NOMy_1/T2; + }}} + } + else { +#pragma omp parallel for shared (A, D2, dimX, dimY) private(i, j, i1, j1, i2, j2, NOMx_1,NOMy_1,NOMx_0,denom1,denom2,T2,index) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + index = j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + + /* Forward-backward differences */ + NOMx_1 = A[j1*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[j*dimX + i1] - A[index]; /* y+ */ + NOMx_0 = A[index] - A[j2*dimX + i]; /* x- */ + /*NOMy_0 = A[(i)*dimY + j] - A[(i)*dimY + j2]; */ /* y- */ + + denom1 = NOMy_1*NOMy_1; + denom2 = 0.5f*(signLLT(NOMx_1) + signLLT(NOMx_0))*(MIN(fabs(NOMx_1),fabs(NOMx_0))); + denom2 = denom2*denom2; + T2 = sqrtf(denom1 + denom2 + EPS_ROF); + D2[index] = NOMy_1/T2; + }} + } + return *D2; +} + +/* calculate differences 3 */ +float D3_func_ROF(float *A, float *D3, long dimX, long dimY, long dimZ) +{ + float NOMx_1, NOMy_1, NOMx_0, NOMy_0, NOMz_1, denom1, denom2, denom3, T3; + long index,i,j,k,i1,i2,k1,j1,j2,k2; + +#pragma omp parallel for shared (A, D3, dimX, dimY, dimZ) private(index, i, j, k, i1, j1, k1, i2, j2, k2, NOMx_1, NOMy_1, NOMy_0, NOMx_0, NOMz_1, denom1, denom2, denom3, T3) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /* Forward-backward differences */ + NOMx_1 = A[(dimX*dimY)*k + (j1)*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[(dimX*dimY)*k + (j)*dimX + i1] - A[index]; /* y+ */ + NOMy_0 = A[index] - A[(dimX*dimY)*k + (j)*dimX + i2]; /* y- */ + NOMx_0 = A[index] - A[(dimX*dimY)*k + (j2)*dimX + i]; /* x- */ + NOMz_1 = A[(dimX*dimY)*k1 + j*dimX + i] - A[index]; /* z+ */ + /*NOMz_0 = A[(dimX*dimY)*k + (i)*dimY + j] - A[(dimX*dimY)*k2 + (i)*dimY + j]; */ /* z- */ + + denom1 = NOMz_1*NOMz_1; + denom2 = 0.5f*(signLLT(NOMx_1) + signLLT(NOMx_0))*(MIN(fabs(NOMx_1),fabs(NOMx_0))); + denom2 = denom2*denom2; + denom3 = 0.5f*(signLLT(NOMy_1) + signLLT(NOMy_0))*(MIN(fabs(NOMy_1),fabs(NOMy_0))); + denom3 = denom3*denom3; + T3 = sqrtf(denom1 + denom2 + denom3 + EPS_ROF); + D3[index] = NOMz_1/T3; + }}} + return *D3; +} + +/*************************************************************************/ +/**********************ROF-LLT-related functions *************************/ +/*************************************************************************/ + +float Update2D_LLT_ROF(float *U0, float *U, float *D1_LLT, float *D2_LLT, float *D1_ROF, float *D2_ROF, float lambdaROF, float lambdaLLT, float tau, long dimX, long dimY, long dimZ) +{ + long i, j, index, i_p, i_m, j_m, j_p; + float div, laplc, dxx, dyy, dv1, dv2; +#pragma omp parallel for shared(U,U0) private(i, j, index, i_p, i_m, j_m, j_p, laplc, div, dxx, dyy, dv1, dv2) + for (i = 0; i<dimX; i++) { + for (j = 0; j<dimY; j++) { + index = j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i_p = i + 1; if (i_p == dimX) i_p = i - 1; + i_m = i - 1; if (i_m < 0) i_m = i + 1; + j_p = j + 1; if (j_p == dimY) j_p = j - 1; + j_m = j - 1; if (j_m < 0) j_m = j + 1; + + /*LLT-related part*/ + dxx = D1_LLT[j*dimX+i_p] - 2.0f*D1_LLT[index] + D1_LLT[j*dimX+i_m]; + dyy = D2_LLT[j_p*dimX+i] - 2.0f*D2_LLT[index] + D2_LLT[j_m*dimX+i]; + laplc = dxx + dyy; /*build Laplacian*/ + + /*ROF-related part*/ + dv1 = D1_ROF[index] - D1_ROF[j_m*dimX + i]; + dv2 = D2_ROF[index] - D2_ROF[j*dimX + i_m]; + div = dv1 + dv2; /*build Divirgent*/ + + /*combine all into one cost function to minimise */ + U[index] += tau*(2.0f*lambdaROF*(div) - lambdaLLT*(laplc) - (U[index] - U0[index])); + } + } + return *U; +} + +float Update3D_LLT_ROF(float *U0, float *U, float *D1_LLT, float *D2_LLT, float *D3_LLT, float *D1_ROF, float *D2_ROF, float *D3_ROF, float lambdaROF, float lambdaLLT, float tau, long dimX, long dimY, long dimZ) +{ + long i, j, k, i_p, i_m, j_m, j_p, k_p, k_m, index; + float div, laplc, dxx, dyy, dzz, dv1, dv2, dv3; +#pragma omp parallel for shared(U,U0) private(i, j, k, index, i_p, i_m, j_m, j_p, k_p, k_m, laplc, div, dxx, dyy, dzz, dv1, dv2, dv3) + for (i = 0; i<dimX; i++) { + for (j = 0; j<dimY; j++) { + for (k = 0; k<dimZ; k++) { + /* symmetric boundary conditions (Neuman) */ + i_p = i + 1; if (i_p == dimX) i_p = i - 1; + i_m = i - 1; if (i_m < 0) i_m = i + 1; + j_p = j + 1; if (j_p == dimY) j_p = j - 1; + j_m = j - 1; if (j_m < 0) j_m = j + 1; + k_p = k + 1; if (k_p == dimZ) k_p = k - 1; + k_m = k - 1; if (k_m < 0) k_m = k + 1; + + index = (dimX*dimY)*k + j*dimX+i; + + /*LLT-related part*/ + dxx = D1_LLT[(dimX*dimY)*k + j*dimX+i_p] - 2.0f*D1_LLT[index] + D1_LLT[(dimX*dimY)*k + j*dimX+i_m]; + dyy = D2_LLT[(dimX*dimY)*k + j_p*dimX+i] - 2.0f*D2_LLT[index] + D2_LLT[(dimX*dimY)*k + j_m*dimX+i]; + dzz = D3_LLT[(dimX*dimY)*k_p + j*dimX+i] - 2.0f*D3_LLT[index] + D3_LLT[(dimX*dimY)*k_m + j*dimX+i]; + laplc = dxx + dyy + dzz; /*build Laplacian*/ + + /*ROF-related part*/ + dv1 = D1_ROF[index] - D1_ROF[(dimX*dimY)*k + j_m*dimX+i]; + dv2 = D2_ROF[index] - D2_ROF[(dimX*dimY)*k + j*dimX+i_m]; + dv3 = D3_ROF[index] - D3_ROF[(dimX*dimY)*k_m + j*dimX+i]; + div = dv1 + dv2 + dv3; /*build Divirgent*/ + + /*combine all into one cost function to minimise */ + U[index] += tau*(2.0f*lambdaROF*(div) - lambdaLLT*(laplc) - (U[index] - U0[index])); + } + } + } + return *U; +} + diff --git a/src/Core/regularisers_CPU/LLT_ROF_core.h b/src/Core/regularisers_CPU/LLT_ROF_core.h new file mode 100644 index 0000000..8e6591e --- /dev/null +++ b/src/Core/regularisers_CPU/LLT_ROF_core.h @@ -0,0 +1,65 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + +/* C-OMP implementation of Lysaker, Lundervold and Tai (LLT) model [1] combined with Rudin-Osher-Fatemi [2] TV regularisation penalty. + * +* This penalty can deliver visually pleasant piecewise-smooth recovery if regularisation parameters are selected well. +* The rule of thumb for selection is to start with lambdaLLT = 0 (just the ROF-TV model) and then proceed to increase +* lambdaLLT starting with smaller values. +* +* Input Parameters: +* 1. U0 - original noise image/volume +* 2. lambdaROF - ROF-related regularisation parameter +* 3. lambdaLLT - LLT-related regularisation parameter +* 4. tau - time-marching step +* 5. iter - iterations number (for both models) +* +* Output: +* Filtered/regularised image +* +* References: +* [1] Lysaker, M., Lundervold, A. and Tai, X.C., 2003. Noise removal using fourth-order partial differential equation with applications to medical magnetic resonance images in space and time. IEEE Transactions on image processing, 12(12), pp.1579-1590. +* [2] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" +*/ + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float LLT_ROF_CPU_main(float *Input, float *Output, float lambdaROF, float lambdaLLT, int iterationsNumb, float tau, int dimX, int dimY, int dimZ); + +CCPI_EXPORT float der2D_LLT(float *U, float *D1, float *D2, long dimX, long dimY, long dimZ); +CCPI_EXPORT float der3D_LLT(float *U, float *D1, float *D2, float *D3, long dimX, long dimY, long dimZ); + +CCPI_EXPORT float D1_func_ROF(float *A, float *D1, long dimX, long dimY, long dimZ); +CCPI_EXPORT float D2_func_ROF(float *A, float *D2, long dimX, long dimY, long dimZ); +CCPI_EXPORT float D3_func_ROF(float *A, float *D3, long dimX, long dimY, long dimZ); + +CCPI_EXPORT float Update2D_LLT_ROF(float *U0, float *U, float *D1_LLT, float *D2_LLT, float *D1_ROF, float *D2_ROF, float lambdaROF, float lambdaLLT, float tau, long dimX, long dimY, long dimZ); +CCPI_EXPORT float Update3D_LLT_ROF(float *U0, float *U, float *D1_LLT, float *D2_LLT, float *D3_LLT, float *D1_ROF, float *D2_ROF, float *D3_ROF, float lambdaROF, float lambdaLLT, float tau, long dimX, long dimY, long dimZ); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/Nonlocal_TV_core.c b/src/Core/regularisers_CPU/Nonlocal_TV_core.c new file mode 100644 index 0000000..c4c9118 --- /dev/null +++ b/src/Core/regularisers_CPU/Nonlocal_TV_core.c @@ -0,0 +1,173 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC and Diamond Light Source Ltd. + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * Copyright 2018 Diamond Light Source Ltd. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "Nonlocal_TV_core.h" + +/* C-OMP implementation of non-local regulariser + * Weights and associated indices must be given as an input. + * Gauss-Seidel fixed point iteration requires ~ 3 iterations, so the main effort + * goes in pre-calculation of weights and selection of patches + * + * + * Input Parameters: + * 1. 2D/3D grayscale image/volume + * 2. AR_i - indeces of i neighbours + * 3. AR_j - indeces of j neighbours + * 4. AR_k - indeces of k neighbours (0 - for 2D case) + * 5. Weights_ij(k) - associated weights + * 6. regularisation parameter + * 7. iterations number + + * Output: + * 1. denoised image/volume + * Elmoataz, Abderrahim, Olivier Lezoray, and Sébastien Bougleux. "Nonlocal discrete regularization on weighted graphs: a framework for image and manifold processing." IEEE Trans. Image Processing 17, no. 7 (2008): 1047-1060. + + */ +/*****************************************************************************/ + +float Nonlocal_TV_CPU_main(float *A_orig, float *Output, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, int dimX, int dimY, int dimZ, int NumNeighb, float lambdaReg, int IterNumb) +{ + + long i, j, k; + int iter; + lambdaReg = 1.0f/lambdaReg; + + /*****2D INPUT *****/ + if (dimZ == 0) { + copyIm(A_orig, Output, (long)(dimX), (long)(dimY), 1l); + /* for each pixel store indeces of the most similar neighbours (patches) */ + for(iter=0; iter<IterNumb; iter++) { +#pragma omp parallel for shared (A_orig, Output, Weights, H_i, H_j, iter) private(i,j) + for(i=0; i<(long)(dimX); i++) { + for(j=0; j<(long)(dimY); j++) { + /*NLM_H1_2D(Output, A_orig, H_i, H_j, Weights, i, j, (long)(dimX), (long)(dimY), NumNeighb, lambdaReg);*/ /* NLM - H1 penalty */ + NLM_TV_2D(Output, A_orig, H_i, H_j, Weights, i, j, (long)(dimX), (long)(dimY), NumNeighb, lambdaReg); /* NLM - TV penalty */ + }} + } + } + else { + /*****3D INPUT *****/ + copyIm(A_orig, Output, (long)(dimX), (long)(dimY), (long)(dimZ)); + /* for each pixel store indeces of the most similar neighbours (patches) */ + for(iter=0; iter<IterNumb; iter++) { +#pragma omp parallel for shared (A_orig, Output, Weights, H_i, H_j, H_k, iter) private(i,j,k) + for(i=0; i<(long)(dimX); i++) { + for(j=0; j<(long)(dimY); j++) { + for(k=0; k<(long)(dimZ); k++) { + /* NLM_H1_3D(Output, A_orig, H_i, H_j, H_k, Weights, i, j, k, dimX, dimY, dimZ, NumNeighb, lambdaReg); */ /* NLM - H1 penalty */ + NLM_TV_3D(Output, A_orig, H_i, H_j, H_k, Weights, i, j, k, (long)(dimX), (long)(dimY), (long)(dimZ), NumNeighb, lambdaReg); /* NLM - TV penalty */ + }}} + } + } + return *Output; +} + +/***********<<<<Main Function for NLM - H1 penalty>>>>**********/ +float NLM_H1_2D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimX, long dimY, int NumNeighb, float lambdaReg) +{ + long x, i1, j1, index, index_m; + float value = 0.0f, normweight = 0.0f; + + index_m = j*dimX+i; + for(x=0; x < NumNeighb; x++) { + index = (dimX*dimY*x) + j*dimX+i; + i1 = H_i[index]; + j1 = H_j[index]; + value += A[j1*dimX+i1]*Weights[index]; + normweight += Weights[index]; + } + A[index_m] = (lambdaReg*A_orig[index_m] + value)/(lambdaReg + normweight); + return *A; +} +/*3D version*/ +float NLM_H1_3D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, long i, long j, long k, long dimX, long dimY, long dimZ, int NumNeighb, float lambdaReg) +{ + long x, i1, j1, k1, index; + float value = 0.0f, normweight = 0.0f; + + for(x=0; x < NumNeighb; x++) { + index = dimX*dimY*dimZ*x + (dimX*dimY*k) + j*dimX+i; + i1 = H_i[index]; + j1 = H_j[index]; + k1 = H_k[index]; + value += A[(dimX*dimY*k1) + j1*dimX+i1]*Weights[index]; + normweight += Weights[index]; + } + A[(dimX*dimY*k) + j*dimX+i] = (lambdaReg*A_orig[(dimX*dimY*k) + j*dimX+i] + value)/(lambdaReg + normweight); + return *A; +} + + +/***********<<<<Main Function for NLM - TV penalty>>>>**********/ +float NLM_TV_2D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimX, long dimY, int NumNeighb, float lambdaReg) +{ + long x, i1, j1, index, index_m; + float value = 0.0f, normweight = 0.0f, NLgrad_magn = 0.0f, NLCoeff; + + index_m = j*dimX+i; + + for(x=0; x < NumNeighb; x++) { + index = (dimX*dimY*x) + j*dimX+i; /*c*/ + i1 = H_i[index]; + j1 = H_j[index]; + NLgrad_magn += powf((A[j1*dimX+i1] - A[index_m]),2)*Weights[index]; + } + + NLgrad_magn = sqrtf(NLgrad_magn); /*Non Local Gradients Magnitude */ + NLCoeff = 2.0f*(1.0f/(NLgrad_magn + EPS)); + + for(x=0; x < NumNeighb; x++) { + index = (dimX*dimY*x) + j*dimX+i; /*c*/ + i1 = H_i[index]; + j1 = H_j[index]; + value += A[j1*dimX+i1]*NLCoeff*Weights[index]; + normweight += Weights[index]*NLCoeff; + } + A[index_m] = (lambdaReg*A_orig[index_m] + value)/(lambdaReg + normweight); + return *A; +} +/*3D version*/ +float NLM_TV_3D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, long i, long j, long k, long dimX, long dimY, long dimZ, int NumNeighb, float lambdaReg) +{ + long x, i1, j1, k1, index; + float value = 0.0f, normweight = 0.0f, NLgrad_magn = 0.0f, NLCoeff; + + for(x=0; x < NumNeighb; x++) { + index = dimX*dimY*dimZ*x + (dimX*dimY*k) + j*dimX+i; + i1 = H_i[index]; + j1 = H_j[index]; + k1 = H_k[index]; + NLgrad_magn += powf((A[(dimX*dimY*k1) + j1*dimX+i1] - A[(dimX*dimY*k1) + j*dimX+i]),2)*Weights[index]; + } + + NLgrad_magn = sqrtf(NLgrad_magn); /*Non Local Gradients Magnitude */ + NLCoeff = 2.0f*(1.0f/(NLgrad_magn + EPS)); + + for(x=0; x < NumNeighb; x++) { + index = dimX*dimY*dimZ*x + (dimX*dimY*k) + j*dimX+i; + i1 = H_i[index]; + j1 = H_j[index]; + k1 = H_k[index]; + value += A[(dimX*dimY*k1) + j1*dimX+i1]*NLCoeff*Weights[index]; + normweight += Weights[index]*NLCoeff; + } + A[(dimX*dimY*k) + j*dimX+i] = (lambdaReg*A_orig[(dimX*dimY*k) + j*dimX+i] + value)/(lambdaReg + normweight); + return *A; +} diff --git a/src/Core/regularisers_CPU/Nonlocal_TV_core.h b/src/Core/regularisers_CPU/Nonlocal_TV_core.h new file mode 100644 index 0000000..6d55101 --- /dev/null +++ b/src/Core/regularisers_CPU/Nonlocal_TV_core.h @@ -0,0 +1,61 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC and Diamond Light Source Ltd. + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * Copyright 2018 Diamond Light Source Ltd. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + +#define EPS 1.0000e-9 + +/* C-OMP implementation of non-local regulariser + * Weights and associated indices must be given as an input. + * Gauss-Seidel fixed point iteration requires ~ 3 iterations, so the main effort + * goes in pre-calculation of weights and selection of patches + * + * + * Input Parameters: + * 1. 2D/3D grayscale image/volume + * 2. AR_i - indeces of i neighbours + * 3. AR_j - indeces of j neighbours + * 4. AR_k - indeces of k neighbours (0 - for 2D case) + * 5. Weights_ij(k) - associated weights + * 6. regularisation parameter + * 7. iterations number + + * Output: + * 1. denoised image/volume + * Elmoataz, Abderrahim, Olivier Lezoray, and Sébastien Bougleux. "Nonlocal discrete regularization on weighted graphs: a framework for image and manifold processing." IEEE Trans. Image Processing 17, no. 7 (2008): 1047-1060. + */ + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float Nonlocal_TV_CPU_main(float *A_orig, float *Output, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, int dimX, int dimY, int dimZ, int NumNeighb, float lambdaReg, int IterNumb); +CCPI_EXPORT float NLM_H1_2D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimX, long dimY, int NumNeighb, float lambdaReg); +CCPI_EXPORT float NLM_TV_2D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimX, long dimY, int NumNeighb, float lambdaReg); +CCPI_EXPORT float NLM_H1_3D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, long i, long j, long k, long dimX, long dimY, long dimZ, int NumNeighb, float lambdaReg); +CCPI_EXPORT float NLM_TV_3D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, long i, long j, long k, long dimX, long dimY, long dimZ, int NumNeighb, float lambdaReg); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/PatchSelect_core.c b/src/Core/regularisers_CPU/PatchSelect_core.c new file mode 100644 index 0000000..cf5cdc7 --- /dev/null +++ b/src/Core/regularisers_CPU/PatchSelect_core.c @@ -0,0 +1,345 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC and Diamond Light Source Ltd. + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * Copyright 2018 Diamond Light Source Ltd. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "PatchSelect_core.h" + +/* C-OMP implementation of non-local weight pre-calculation for non-local priors + * Weights and associated indices are stored into pre-allocated arrays and passed + * to the regulariser + * + * + * Input Parameters: + * 1. 2D/3D grayscale image/volume + * 2. Searching window (half-size of the main bigger searching window, e.g. 11) + * 3. Similarity window (half-size of the patch window, e.g. 2) + * 4. The number of neighbours to take (the most prominent after sorting neighbours will be taken) + * 5. noise-related parameter to calculate non-local weights + * + * Output [2D]: + * 1. AR_i - indeces of i neighbours + * 2. AR_j - indeces of j neighbours + * 3. Weights_ij - associated weights + * + * Output [3D]: + * 1. AR_i - indeces of i neighbours + * 2. AR_j - indeces of j neighbours + * 3. AR_k - indeces of j neighbours + * 4. Weights_ijk - associated weights + */ + +void swap(float *xp, float *yp) +{ + float temp = *xp; + *xp = *yp; + *yp = temp; +} + +void swapUS(unsigned short *xp, unsigned short *yp) +{ + unsigned short temp = *xp; + *xp = *yp; + *yp = temp; +} +/**************************************************/ + +float PatchSelect_CPU_main(float *A, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, int dimX, int dimY, int dimZ, int SearchWindow, int SimilarWin, int NumNeighb, float h, int switchM) +{ + int counterG; + long i, j, k; + float *Eucl_Vec, h2; + h2 = h*h; + /****************2D INPUT ***************/ + if (dimZ == 0) { + /* generate a 2D Gaussian kernel for NLM procedure */ + Eucl_Vec = (float*) calloc ((2*SimilarWin+1)*(2*SimilarWin+1),sizeof(float)); + counterG = 0; + for(i=-SimilarWin; i<=SimilarWin; i++) { + for(j=-SimilarWin; j<=SimilarWin; j++) { + Eucl_Vec[counterG] = (float)exp(-(pow(((float) i), 2) + pow(((float) j), 2))/(2*SimilarWin*SimilarWin)); + counterG++; + }} /*main neighb loop */ + /* for each pixel store indeces of the most similar neighbours (patches) */ + if (switchM == 1) { +#pragma omp parallel for shared (A, Weights, H_i, H_j) private(i,j) + for(i=0; i<(long)(dimX); i++) { + for(j=0; j<(long)(dimY); j++) { + Indeces2D_p(A, H_i, H_j, Weights, i, j, (long)(dimX), (long)(dimY), Eucl_Vec, NumNeighb, SearchWindow, SimilarWin, h2); + }} + } + else { +#pragma omp parallel for shared (A, Weights, H_i, H_j) private(i,j) + for(i=0; i<(long)(dimX); i++) { + for(j=0; j<(long)(dimY); j++) { + Indeces2D(A, H_i, H_j, Weights, i, j, (long)(dimX), (long)(dimY), Eucl_Vec, NumNeighb, SearchWindow, SimilarWin, h2); + }} + } + } + else { + /****************3D INPUT ***************/ + /* generate a 3D Gaussian kernel for NLM procedure */ + Eucl_Vec = (float*) calloc ((2*SimilarWin+1)*(2*SimilarWin+1)*(2*SimilarWin+1),sizeof(float)); + counterG = 0; + for(i=-SimilarWin; i<=SimilarWin; i++) { + for(j=-SimilarWin; j<=SimilarWin; j++) { + for(k=-SimilarWin; k<=SimilarWin; k++) { + Eucl_Vec[counterG] = (float)exp(-(pow(((float) i), 2) + pow(((float) j), 2) + pow(((float) k), 2))/(2*SimilarWin*SimilarWin*SimilarWin)); + counterG++; + }}} /*main neighb loop */ + + /* for each voxel store indeces of the most similar neighbours (patches) */ + if (switchM == 1) { +#pragma omp parallel for shared (A, Weights, H_i, H_j, H_k) private(i,j,k) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + Indeces3D(A, H_i, H_j, H_k, Weights, j, i, (k), (dimX), (dimY), (dimZ), Eucl_Vec, NumNeighb, SearchWindow, SimilarWin, h2); + }}} + } + else { +#pragma omp parallel for shared (A, Weights, H_i, H_j, H_k) private(i,j,k) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + Indeces3D(A, H_i, H_j, H_k, Weights, (i), (j), (k), (dimX), (dimY), (dimZ), Eucl_Vec, NumNeighb, SearchWindow, SimilarWin, h2); + }}} + } + } + free(Eucl_Vec); + return 1; +} + +float Indeces2D(float *Aorig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimX, long dimY, float *Eucl_Vec, int NumNeighb, int SearchWindow, int SimilarWin, float h2) +{ + long i1, j1, i_m, j_m, i_c, j_c, i2, j2, i3, j3, counter, x, y, index, sizeWin_tot, counterG; + float *Weight_Vec, normsum; + unsigned short *ind_i, *ind_j; + + sizeWin_tot = (2*SearchWindow + 1)*(2*SearchWindow + 1); + + Weight_Vec = (float*) calloc(sizeWin_tot, sizeof(float)); + ind_i = (unsigned short*) calloc(sizeWin_tot, sizeof(unsigned short)); + ind_j = (unsigned short*) calloc(sizeWin_tot, sizeof(unsigned short)); + + counter = 0; + for(i_m=-SearchWindow; i_m<=SearchWindow; i_m++) { + for(j_m=-SearchWindow; j_m<=SearchWindow; j_m++) { + i1 = i+i_m; + j1 = j+j_m; + if (((i1 >= 0) && (i1 < dimX)) && ((j1 >= 0) && (j1 < dimY))) { + normsum = 0.0f; counterG = 0; + for(i_c=-SimilarWin; i_c<=SimilarWin; i_c++) { + for(j_c=-SimilarWin; j_c<=SimilarWin; j_c++) { + i2 = i1 + i_c; + j2 = j1 + j_c; + i3 = i + i_c; + j3 = j + j_c; + if (((i2 >= 0) && (i2 < dimX)) && ((j2 >= 0) && (j2 < dimY))) { + if (((i3 >= 0) && (i3 < dimX)) && ((j3 >= 0) && (j3 < dimY))) { + normsum += Eucl_Vec[counterG]*pow(Aorig[j3*dimX + (i3)] - Aorig[j2*dimX + (i2)], 2); + counterG++; + }} + + }} + /* writing temporarily into vectors */ + if (normsum > EPS) { + Weight_Vec[counter] = expf(-normsum/h2); + ind_i[counter] = i1; + ind_j[counter] = j1; + counter++; + } + } + }} + /* do sorting to choose the most prominent weights [HIGH to LOW] */ + /* and re-arrange indeces accordingly */ + for (x = 0; x < counter-1; x++) { + for (y = 0; y < counter-x-1; y++) { + if (Weight_Vec[y] < Weight_Vec[y+1]) { + swap(&Weight_Vec[y], &Weight_Vec[y+1]); + swapUS(&ind_i[y], &ind_i[y+1]); + swapUS(&ind_j[y], &ind_j[y+1]); + } + } + } + /*sorting loop finished*/ + /*now select the NumNeighb more prominent weights and store into pre-allocated arrays */ + for(x=0; x < NumNeighb; x++) { + index = (dimX*dimY*x) + j*dimX+i; + H_i[index] = ind_i[x]; + H_j[index] = ind_j[x]; + Weights[index] = Weight_Vec[x]; + } + free(ind_i); + free(ind_j); + free(Weight_Vec); + return 1; +} +float Indeces2D_p(float *Aorig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimX, long dimY, float *Eucl_Vec, int NumNeighb, int SearchWindow, int SimilarWin, float h2) +{ + long i1, j1, i_m, j_m, i_c, j_c, i2, j2, i3, j3, counter, x, y, index, sizeWin_tot, counterG; + float *Weight_Vec, normsum; + unsigned short *ind_i, *ind_j; + + sizeWin_tot = (2*SearchWindow + 1)*(2*SearchWindow + 1); + + Weight_Vec = (float*) calloc(sizeWin_tot, sizeof(float)); + ind_i = (unsigned short*) calloc(sizeWin_tot, sizeof(unsigned short)); + ind_j = (unsigned short*) calloc(sizeWin_tot, sizeof(unsigned short)); + + counter = 0; + for(i_m=-SearchWindow; i_m<=SearchWindow; i_m++) { + for(j_m=-SearchWindow; j_m<=SearchWindow; j_m++) { + i1 = i+i_m; + j1 = j+j_m; + if (((i1 >= 0) && (i1 < dimX)) && ((j1 >= 0) && (j1 < dimY))) { + normsum = 0.0f; counterG = 0; + for(i_c=-SimilarWin; i_c<=SimilarWin; i_c++) { + for(j_c=-SimilarWin; j_c<=SimilarWin; j_c++) { + i2 = i1 + i_c; + j2 = j1 + j_c; + i3 = i + i_c; + j3 = j + j_c; + if (((i2 >= 0) && (i2 < dimX)) && ((j2 >= 0) && (j2 < dimY))) { + if (((i3 >= 0) && (i3 < dimX)) && ((j3 >= 0) && (j3 < dimY))) { + //normsum += Eucl_Vec[counterG]*pow(Aorig[j3*dimX + (i3)] - Aorig[j2*dimX + (i2)], 2); + normsum += Eucl_Vec[counterG]*pow(Aorig[i3*dimY + (j3)] - Aorig[i2*dimY + (j2)], 2); + counterG++; + }} + + }} + /* writing temporarily into vectors */ + if (normsum > EPS) { + Weight_Vec[counter] = expf(-normsum/h2); + ind_i[counter] = i1; + ind_j[counter] = j1; + counter++; + } + } + }} + /* do sorting to choose the most prominent weights [HIGH to LOW] */ + /* and re-arrange indeces accordingly */ + for (x = 0; x < counter-1; x++) { + for (y = 0; y < counter-x-1; y++) { + if (Weight_Vec[y] < Weight_Vec[y+1]) { + swap(&Weight_Vec[y], &Weight_Vec[y+1]); + swapUS(&ind_i[y], &ind_i[y+1]); + swapUS(&ind_j[y], &ind_j[y+1]); + } + } + } + /*sorting loop finished*/ + + /*now select the NumNeighb more prominent weights and store into pre-allocated arrays */ + for(x=0; x < NumNeighb; x++) { + index = (dimX*dimY*x) + i*dimY+j; + H_i[index] = ind_i[x]; + H_j[index] = ind_j[x]; + Weights[index] = Weight_Vec[x]; + } + free(ind_i); + free(ind_j); + free(Weight_Vec); + return 1; +} + +float Indeces3D(float *Aorig, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, long i, long j, long k, long dimY, long dimX, long dimZ, float *Eucl_Vec, int NumNeighb, int SearchWindow, int SimilarWin, float h2) +{ + long i1, j1, k1, i_m, j_m, k_m, i_c, j_c, k_c, i2, j2, k2, i3, j3, k3, counter, x, y, index, sizeWin_tot, counterG; + float *Weight_Vec, normsum, temp; + unsigned short *ind_i, *ind_j, *ind_k, temp_i, temp_j, temp_k; + + sizeWin_tot = (2*SearchWindow + 1)*(2*SearchWindow + 1)*(2*SearchWindow + 1); + + Weight_Vec = (float*) calloc(sizeWin_tot, sizeof(float)); + ind_i = (unsigned short*) calloc(sizeWin_tot, sizeof(unsigned short)); + ind_j = (unsigned short*) calloc(sizeWin_tot, sizeof(unsigned short)); + ind_k = (unsigned short*) calloc(sizeWin_tot, sizeof(unsigned short)); + + counter = 0l; + for(i_m=-SearchWindow; i_m<=SearchWindow; i_m++) { + for(j_m=-SearchWindow; j_m<=SearchWindow; j_m++) { + for(k_m=-SearchWindow; k_m<=SearchWindow; k_m++) { + k1 = k+k_m; + i1 = i+i_m; + j1 = j+j_m; + if (((i1 >= 0) && (i1 < dimX)) && ((j1 >= 0) && (j1 < dimY)) && ((k1 >= 0) && (k1 < dimZ))) { + normsum = 0.0f; counterG = 0l; + for(i_c=-SimilarWin; i_c<=SimilarWin; i_c++) { + for(j_c=-SimilarWin; j_c<=SimilarWin; j_c++) { + for(k_c=-SimilarWin; k_c<=SimilarWin; k_c++) { + i2 = i1 + i_c; + j2 = j1 + j_c; + k2 = k1 + k_c; + i3 = i + i_c; + j3 = j + j_c; + k3 = k + k_c; + if (((i2 >= 0) && (i2 < dimX)) && ((j2 >= 0) && (j2 < dimY)) && ((k2 >= 0) && (k2 < dimZ))) { + if (((i3 >= 0) && (i3 < dimX)) && ((j3 >= 0) && (j3 < dimY)) && ((k3 >= 0) && (k3 < dimZ))) { + normsum += Eucl_Vec[counterG]*pow(Aorig[(dimX*dimY*k3) + j3*dimX + (i3)] - Aorig[(dimX*dimY*k2) + j2*dimX + (i2)], 2); + counterG++; + }} + }}} + /* writing temporarily into vectors */ + if (normsum > EPS) { + Weight_Vec[counter] = expf(-normsum/h2); + ind_i[counter] = i1; + ind_j[counter] = j1; + ind_k[counter] = k1; + counter ++; + } + } + }}} + /* do sorting to choose the most prominent weights [HIGH to LOW] */ + /* and re-arrange indeces accordingly */ + for (x = 0; x < counter; x++) { + for (y = 0; y < counter; y++) { + if (Weight_Vec[y] < Weight_Vec[x]) { + temp = Weight_Vec[y+1]; + temp_i = ind_i[y+1]; + temp_j = ind_j[y+1]; + temp_k = ind_k[y+1]; + Weight_Vec[y+1] = Weight_Vec[y]; + Weight_Vec[y] = temp; + ind_i[y+1] = ind_i[y]; + ind_i[y] = temp_i; + ind_j[y+1] = ind_j[y]; + ind_j[y] = temp_j; + ind_k[y+1] = ind_k[y]; + ind_k[y] = temp_k; + }}} + /*sorting loop finished*/ + + /*now select the NumNeighb more prominent weights and store into arrays */ + for(x=0; x < NumNeighb; x++) { + index = dimX*dimY*dimZ*x + (dimX*dimY*k) + j*dimX+i; + + H_i[index] = ind_i[x]; + H_j[index] = ind_j[x]; + H_k[index] = ind_k[x]; + + Weights[index] = Weight_Vec[x]; + } + + free(ind_i); + free(ind_j); + free(ind_k); + free(Weight_Vec); + return 1; +} + diff --git a/src/Core/regularisers_CPU/PatchSelect_core.h b/src/Core/regularisers_CPU/PatchSelect_core.h new file mode 100644 index 0000000..ddaa428 --- /dev/null +++ b/src/Core/regularisers_CPU/PatchSelect_core.h @@ -0,0 +1,63 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC and Diamond Light Source Ltd. + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * Copyright 2018 Diamond Light Source Ltd. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" +#define EPS 1.0000e-12 + +/* C-OMP implementation of non-local weight pre-calculation for non-local priors + * Weights and associated indices are stored into pre-allocated arrays and passed + * to the regulariser + * + * + * Input Parameters: + * 1. 2D/3D grayscale image/volume + * 2. Searching window (half-size of the main bigger searching window, e.g. 11) + * 3. Similarity window (half-size of the patch window, e.g. 2) + * 4. The number of neighbours to take (the most prominent after sorting neighbours will be taken) + * 5. noise-related parameter to calculate non-local weights + * + * Output [2D]: + * 1. AR_i - indeces of i neighbours + * 2. AR_j - indeces of j neighbours + * 3. Weights_ij - associated weights + * + * Output [3D]: + * 1. AR_i - indeces of i neighbours + * 2. AR_j - indeces of j neighbours + * 3. AR_k - indeces of j neighbours + * 4. Weights_ijk - associated weights + */ +/*****************************************************************************/ +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float PatchSelect_CPU_main(float *A, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, int dimX, int dimY, int dimZ, int SearchWindow, int SimilarWin, int NumNeighb, float h, int switchM); +CCPI_EXPORT float Indeces2D(float *Aorig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimX, long dimY, float *Eucl_Vec, int NumNeighb, int SearchWindow, int SimilarWin, float h2); +CCPI_EXPORT float Indeces2D_p(float *Aorig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimX, long dimY, float *Eucl_Vec, int NumNeighb, int SearchWindow, int SimilarWin, float h2); +CCPI_EXPORT float Indeces3D(float *Aorig, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, long i, long j, long k, long dimY, long dimX, long dimZ, float *Eucl_Vec, int NumNeighb, int SearchWindow, int SimilarWin, float h2); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/ROF_TV_core.c b/src/Core/regularisers_CPU/ROF_TV_core.c new file mode 100644 index 0000000..1858442 --- /dev/null +++ b/src/Core/regularisers_CPU/ROF_TV_core.c @@ -0,0 +1,289 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "ROF_TV_core.h" + +#define EPS 1.0e-12 +#define MAX(x, y) (((x) > (y)) ? (x) : (y)) +#define MIN(x, y) (((x) < (y)) ? (x) : (y)) + +/*sign function*/ +int sign(float x) { + return (x > 0) - (x < 0); +} + + +/* C-OMP implementation of ROF-TV denoising/regularization model [1] (2D/3D case) + * + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. lambda - regularization parameter [REQUIRED] + * 3. tau - marching step for explicit scheme, ~1 is recommended [REQUIRED] + * 4. Number of iterations, for explicit scheme >= 150 is recommended [REQUIRED] + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" + */ + +/* Running iterations of TV-ROF function */ +float TV_ROF_CPU_main(float *Input, float *Output, float lambdaPar, int iterationsNumb, float tau, int dimX, int dimY, int dimZ) +{ + float *D1, *D2, *D3; + int i; + long DimTotal; + DimTotal = (long)(dimX*dimY*dimZ); + + D1 = calloc(DimTotal, sizeof(float)); + D2 = calloc(DimTotal, sizeof(float)); + D3 = calloc(DimTotal, sizeof(float)); + + /* copy into output */ + copyIm(Input, Output, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* start TV iterations */ + for(i=0; i < iterationsNumb; i++) { + /* calculate differences */ + D1_func(Output, D1, (long)(dimX), (long)(dimY), (long)(dimZ)); + D2_func(Output, D2, (long)(dimX), (long)(dimY), (long)(dimZ)); + if (dimZ > 1) D3_func(Output, D3, (long)(dimX), (long)(dimY), (long)(dimZ)); + TV_kernel(D1, D2, D3, Output, Input, lambdaPar, tau, (long)(dimX), (long)(dimY), (long)(dimZ)); + } + free(D1);free(D2); free(D3); + return *Output; +} + +/* calculate differences 1 */ +float D1_func(float *A, float *D1, long dimX, long dimY, long dimZ) +{ + float NOMx_1, NOMy_1, NOMy_0, NOMz_1, NOMz_0, denom1, denom2,denom3, T1; + long i,j,k,i1,i2,k1,j1,j2,k2,index; + + if (dimZ > 1) { +#pragma omp parallel for shared (A, D1, dimX, dimY, dimZ) private(index, i, j, k, i1, j1, k1, i2, j2, k2, NOMx_1,NOMy_1,NOMy_0,NOMz_1,NOMz_0,denom1,denom2,denom3,T1) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /* Forward-backward differences */ + NOMx_1 = A[(dimX*dimY)*k + j1*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[(dimX*dimY)*k + j*dimX + i1] - A[index]; /* y+ */ + /*NOMx_0 = (A[(i)*dimY + j] - A[(i2)*dimY + j]); */ /* x- */ + NOMy_0 = A[index] - A[(dimX*dimY)*k + j*dimX + i2]; /* y- */ + + NOMz_1 = A[(dimX*dimY)*k1 + j*dimX + i] - A[index]; /* z+ */ + NOMz_0 = A[index] - A[(dimX*dimY)*k2 + j*dimX + i]; /* z- */ + + + denom1 = NOMx_1*NOMx_1; + denom2 = 0.5f*(sign(NOMy_1) + sign(NOMy_0))*(MIN(fabs(NOMy_1),fabs(NOMy_0))); + denom2 = denom2*denom2; + denom3 = 0.5f*(sign(NOMz_1) + sign(NOMz_0))*(MIN(fabs(NOMz_1),fabs(NOMz_0))); + denom3 = denom3*denom3; + T1 = sqrt(denom1 + denom2 + denom3 + EPS); + D1[index] = NOMx_1/T1; + }}} + } + else { +#pragma omp parallel for shared (A, D1, dimX, dimY) private(i, j, i1, j1, i2, j2,NOMx_1,NOMy_1,NOMy_0,denom1,denom2,T1,index) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + index = j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + + /* Forward-backward differences */ + NOMx_1 = A[j1*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[j*dimX + i1] - A[index]; /* y+ */ + /*NOMx_0 = (A[(i)*dimY + j] - A[(i2)*dimY + j]); */ /* x- */ + NOMy_0 = A[index] - A[(j)*dimX + i2]; /* y- */ + + denom1 = NOMx_1*NOMx_1; + denom2 = 0.5f*(sign(NOMy_1) + sign(NOMy_0))*(MIN(fabs(NOMy_1),fabs(NOMy_0))); + denom2 = denom2*denom2; + T1 = sqrtf(denom1 + denom2 + EPS); + D1[index] = NOMx_1/T1; + }} + } + return *D1; +} +/* calculate differences 2 */ +float D2_func(float *A, float *D2, long dimX, long dimY, long dimZ) +{ + float NOMx_1, NOMy_1, NOMx_0, NOMz_1, NOMz_0, denom1, denom2, denom3, T2; + long i,j,k,i1,i2,k1,j1,j2,k2,index; + + if (dimZ > 1) { +#pragma omp parallel for shared (A, D2, dimX, dimY, dimZ) private(index, i, j, k, i1, j1, k1, i2, j2, k2, NOMx_1, NOMy_1, NOMx_0, NOMz_1, NOMz_0, denom1, denom2, denom3, T2) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /* Forward-backward differences */ + NOMx_1 = A[(dimX*dimY)*k + (j1)*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[(dimX*dimY)*k + (j)*dimX + i1] - A[index]; /* y+ */ + NOMx_0 = A[index] - A[(dimX*dimY)*k + (j2)*dimX + i]; /* x- */ + NOMz_1 = A[(dimX*dimY)*k1 + j*dimX + i] - A[index]; /* z+ */ + NOMz_0 = A[index] - A[(dimX*dimY)*k2 + (j)*dimX + i]; /* z- */ + + + denom1 = NOMy_1*NOMy_1; + denom2 = 0.5f*(sign(NOMx_1) + sign(NOMx_0))*(MIN(fabs(NOMx_1),fabs(NOMx_0))); + denom2 = denom2*denom2; + denom3 = 0.5f*(sign(NOMz_1) + sign(NOMz_0))*(MIN(fabs(NOMz_1),fabs(NOMz_0))); + denom3 = denom3*denom3; + T2 = sqrtf(denom1 + denom2 + denom3 + EPS); + D2[index] = NOMy_1/T2; + }}} + } + else { +#pragma omp parallel for shared (A, D2, dimX, dimY) private(i, j, i1, j1, i2, j2, NOMx_1,NOMy_1,NOMx_0,denom1,denom2,T2,index) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + index = j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + + /* Forward-backward differences */ + NOMx_1 = A[j1*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[j*dimX + i1] - A[index]; /* y+ */ + NOMx_0 = A[index] - A[j2*dimX + i]; /* x- */ + /*NOMy_0 = A[(i)*dimY + j] - A[(i)*dimY + j2]; */ /* y- */ + + denom1 = NOMy_1*NOMy_1; + denom2 = 0.5f*(sign(NOMx_1) + sign(NOMx_0))*(MIN(fabs(NOMx_1),fabs(NOMx_0))); + denom2 = denom2*denom2; + T2 = sqrtf(denom1 + denom2 + EPS); + D2[index] = NOMy_1/T2; + }} + } + return *D2; +} + +/* calculate differences 3 */ +float D3_func(float *A, float *D3, long dimX, long dimY, long dimZ) +{ + float NOMx_1, NOMy_1, NOMx_0, NOMy_0, NOMz_1, denom1, denom2, denom3, T3; + long index,i,j,k,i1,i2,k1,j1,j2,k2; + +#pragma omp parallel for shared (A, D3, dimX, dimY, dimZ) private(index, i, j, k, i1, j1, k1, i2, j2, k2, NOMx_1, NOMy_1, NOMy_0, NOMx_0, NOMz_1, denom1, denom2, denom3, T3) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /* Forward-backward differences */ + NOMx_1 = A[(dimX*dimY)*k + (j1)*dimX + i] - A[index]; /* x+ */ + NOMy_1 = A[(dimX*dimY)*k + (j)*dimX + i1] - A[index]; /* y+ */ + NOMy_0 = A[index] - A[(dimX*dimY)*k + (j)*dimX + i2]; /* y- */ + NOMx_0 = A[index] - A[(dimX*dimY)*k + (j2)*dimX + i]; /* x- */ + NOMz_1 = A[(dimX*dimY)*k1 + j*dimX + i] - A[index]; /* z+ */ + /*NOMz_0 = A[(dimX*dimY)*k + (i)*dimY + j] - A[(dimX*dimY)*k2 + (i)*dimY + j]; */ /* z- */ + + denom1 = NOMz_1*NOMz_1; + denom2 = 0.5f*(sign(NOMx_1) + sign(NOMx_0))*(MIN(fabs(NOMx_1),fabs(NOMx_0))); + denom2 = denom2*denom2; + denom3 = 0.5f*(sign(NOMy_1) + sign(NOMy_0))*(MIN(fabs(NOMy_1),fabs(NOMy_0))); + denom3 = denom3*denom3; + T3 = sqrtf(denom1 + denom2 + denom3 + EPS); + D3[index] = NOMz_1/T3; + }}} + return *D3; +} + +/* calculate divergence */ +float TV_kernel(float *D1, float *D2, float *D3, float *B, float *A, float lambda, float tau, long dimX, long dimY, long dimZ) +{ + float dv1, dv2, dv3; + long index,i,j,k,i1,i2,k1,j1,j2,k2; + + if (dimZ > 1) { +#pragma omp parallel for shared (D1, D2, D3, B, dimX, dimY, dimZ) private(index, i, j, k, i1, j1, k1, i2, j2, k2, dv1,dv2,dv3) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /*divergence components */ + dv1 = D1[index] - D1[(dimX*dimY)*k + j2*dimX+i]; + dv2 = D2[index] - D2[(dimX*dimY)*k + j*dimX+i2]; + dv3 = D3[index] - D3[(dimX*dimY)*k2 + j*dimX+i]; + + B[index] += tau*(2.0f*lambda*(dv1 + dv2 + dv3) - (B[index] - A[index])); + }}} + } + else { +#pragma omp parallel for shared (D1, D2, B, dimX, dimY) private(index, i, j, i1, j1, i2, j2,dv1,dv2) + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + index = j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + + /* divergence components */ + dv1 = D1[index] - D1[j2*dimX + i]; + dv2 = D2[index] - D2[j*dimX + i2]; + + B[index] += tau*(2.0f*lambda*(dv1 + dv2) - (B[index] - A[index])); + }} + } + return *B; +} diff --git a/src/Core/regularisers_CPU/ROF_TV_core.h b/src/Core/regularisers_CPU/ROF_TV_core.h new file mode 100644 index 0000000..4e320e9 --- /dev/null +++ b/src/Core/regularisers_CPU/ROF_TV_core.h @@ -0,0 +1,57 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + +/* C-OMP implementation of ROF-TV denoising/regularization model [1] (2D/3D case) + * + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. lambda - regularization parameter [REQUIRED] + * 3. Number of iterations, for explicit scheme >= 150 is recommended [REQUIRED] + * 4. tau - marching step for explicit scheme, ~1 is recommended [REQUIRED] + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" + * + * D. Kazantsev, 2016-18 + */ + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float TV_ROF_CPU_main(float *Input, float *Output, float lambdaPar, int iterationsNumb, float tau, int dimX, int dimY, int dimZ); + +CCPI_EXPORT float TV_kernel(float *D1, float *D2, float *D3, float *B, float *A, float lambda, float tau, long dimX, long dimY, long dimZ); +CCPI_EXPORT float D1_func(float *A, float *D1, long dimX, long dimY, long dimZ); +CCPI_EXPORT float D2_func(float *A, float *D2, long dimX, long dimY, long dimZ); +CCPI_EXPORT float D3_func(float *A, float *D3, long dimX, long dimY, long dimZ); +#ifdef __cplusplus +} +#endif
\ No newline at end of file diff --git a/src/Core/regularisers_CPU/SB_TV_core.c b/src/Core/regularisers_CPU/SB_TV_core.c new file mode 100755 index 0000000..769ea67 --- /dev/null +++ b/src/Core/regularisers_CPU/SB_TV_core.c @@ -0,0 +1,368 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "SB_TV_core.h" + +/* C-OMP implementation of Split Bregman - TV denoising-regularisation model (2D/3D) [1] +* +* Input Parameters: +* 1. Noisy image/volume +* 2. lambda - regularisation parameter +* 3. Number of iterations [OPTIONAL parameter] +* 4. eplsilon - tolerance constant [OPTIONAL parameter] +* 5. TV-type: 'iso' or 'l1' [OPTIONAL parameter] +* 6. print information: 0 (off) or 1 (on) [OPTIONAL parameter] +* +* Output: +* 1. Filtered/regularized image +* +* [1]. Goldstein, T. and Osher, S., 2009. The split Bregman method for L1-regularized problems. SIAM journal on imaging sciences, 2(2), pp.323-343. +*/ + +float SB_TV_CPU_main(float *Input, float *Output, float mu, int iter, float epsil, int methodTV, int printM, int dimX, int dimY, int dimZ) +{ + int ll; + long j, DimTotal; + float re, re1, lambda; + int count = 0; + mu = 1.0f/mu; + lambda = 2.0f*mu; + + if (dimZ <= 1) { + /* 2D case */ + float *Output_prev=NULL, *Dx=NULL, *Dy=NULL, *Bx=NULL, *By=NULL; + DimTotal = (long)(dimX*dimY); + + Output_prev = calloc(DimTotal, sizeof(float)); + Dx = calloc(DimTotal, sizeof(float)); + Dy = calloc(DimTotal, sizeof(float)); + Bx = calloc(DimTotal, sizeof(float)); + By = calloc(DimTotal, sizeof(float)); + + copyIm(Input, Output, (long)(dimX), (long)(dimY), 1l); /*initialize */ + + /* begin outer SB iterations */ + for(ll=0; ll<iter; ll++) { + + /* storing old estimate */ + copyIm(Output, Output_prev, (long)(dimX), (long)(dimY), 1l); + + /* perform two GS iterations (normally 2 is enough for the convergence) */ + gauss_seidel2D(Output, Input, Output_prev, Dx, Dy, Bx, By, (long)(dimX), (long)(dimY), lambda, mu); + copyIm(Output, Output_prev, (long)(dimX), (long)(dimY), 1l); + /*GS iteration */ + gauss_seidel2D(Output, Input, Output_prev, Dx, Dy, Bx, By, (long)(dimX), (long)(dimY), lambda, mu); + + /* TV-related step */ + if (methodTV == 1) updDxDy_shrinkAniso2D(Output, Dx, Dy, Bx, By, (long)(dimX), (long)(dimY), lambda); + else updDxDy_shrinkIso2D(Output, Dx, Dy, Bx, By, (long)(dimX), (long)(dimY), lambda); + + /* update for Bregman variables */ + updBxBy2D(Output, Dx, Dy, Bx, By, (long)(dimX), (long)(dimY)); + + /* check early stopping criteria if epsilon not equal zero */ + if (epsil != 0) { + re = 0.0f; re1 = 0.0f; + for(j=0; j<DimTotal; j++) { + re += pow(Output[j] - Output_prev[j],2); + re1 += pow(Output[j],2); + } + re = sqrt(re)/sqrt(re1); + if (re < epsil) count++; + if (count > 4) break; + } + /*printf("%f %i %i \n", re, ll, count); */ + } + if (printM == 1) printf("SB-TV iterations stopped at iteration %i \n", ll); + free(Output_prev); free(Dx); free(Dy); free(Bx); free(By); + } + else { + /* 3D case */ + float *Output_prev=NULL, *Dx=NULL, *Dy=NULL, *Dz=NULL, *Bx=NULL, *By=NULL, *Bz=NULL; + DimTotal = (long)(dimX*dimY*dimZ); + + Output_prev = calloc(DimTotal, sizeof(float)); + Dx = calloc(DimTotal, sizeof(float)); + Dy = calloc(DimTotal, sizeof(float)); + Dz = calloc(DimTotal, sizeof(float)); + Bx = calloc(DimTotal, sizeof(float)); + By = calloc(DimTotal, sizeof(float)); + Bz = calloc(DimTotal, sizeof(float)); + + copyIm(Input, Output, (long)(dimX), (long)(dimY), (long)(dimZ)); /*initialize */ + + /* begin outer SB iterations */ + for(ll=0; ll<iter; ll++) { + + /* storing old estimate */ + copyIm(Output, Output_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* perform two GS iterations (normally 2 is enough for the convergence) */ + gauss_seidel3D(Output, Input, Output_prev, Dx, Dy, Dz, Bx, By, Bz, (long)(dimX), (long)(dimY), (long)(dimZ), lambda, mu); + copyIm(Output, Output_prev, (long)(dimX), (long)(dimY), (long)(dimZ)); + /*GS iteration */ + gauss_seidel3D(Output, Input, Output_prev, Dx, Dy, Dz, Bx, By, Bz, (long)(dimX), (long)(dimY), (long)(dimZ), lambda, mu); + + /* TV-related step */ + if (methodTV == 1) updDxDyDz_shrinkAniso3D(Output, Dx, Dy, Dz, Bx, By, Bz, (long)(dimX), (long)(dimY), (long)(dimZ), lambda); + else updDxDyDz_shrinkIso3D(Output, Dx, Dy, Dz, Bx, By, Bz, (long)(dimX), (long)(dimY), (long)(dimZ), lambda); + + /* update for Bregman variables */ + updBxByBz3D(Output, Dx, Dy, Dz, Bx, By, Bz, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* check early stopping criteria if epsilon not equal zero */ + if (epsil != 0) { + re = 0.0f; re1 = 0.0f; + for(j=0; j<DimTotal; j++) { + re += pow(Output[j] - Output_prev[j],2); + re1 += pow(Output[j],2); + } + re = sqrt(re)/sqrt(re1); + if (re < epsil) count++; + if (count > 4) break; + } + /*printf("%f %i %i \n", re, ll, count); */ + } + if (printM == 1) printf("SB-TV iterations stopped at iteration %i \n", ll); + free(Output_prev); free(Dx); free(Dy); free(Dz); free(Bx); free(By); free(Bz); + } + return *Output; +} + +/********************************************************************/ +/***************************2D Functions*****************************/ +/********************************************************************/ +float gauss_seidel2D(float *U, float *A, float *U_prev, float *Dx, float *Dy, float *Bx, float *By, long dimX, long dimY, float lambda, float mu) +{ + float sum, normConst; + long i,j,i1,i2,j1,j2,index; + normConst = 1.0f/(mu + 4.0f*lambda); + +#pragma omp parallel for shared(U) private(index,i,j,i1,i2,j1,j2,sum) + for(i=0; i<dimX; i++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + index = j*dimX+i; + + sum = Dx[j*dimX+i2] - Dx[index] + Dy[j2*dimX+i] - Dy[index] - Bx[j*dimX+i2] + Bx[index] - By[j2*dimX+i] + By[index]; + sum += U_prev[j*dimX+i1] + U_prev[j*dimX+i2] + U_prev[j1*dimX+i] + U_prev[j2*dimX+i]; + sum *= lambda; + sum += mu*A[index]; + U[index] = normConst*sum; + }} + return *U; +} + +float updDxDy_shrinkAniso2D(float *U, float *Dx, float *Dy, float *Bx, float *By, long dimX, long dimY, float lambda) +{ + long i,j,i1,j1,index; + float val1, val11, val2, val22, denom_lam; + denom_lam = 1.0f/lambda; +#pragma omp parallel for shared(U,denom_lam) private(index,i,j,i1,j1,val1,val11,val2,val22) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + index = j*dimX+i; + + val1 = (U[j*dimX+i1] - U[index]) + Bx[index]; + val2 = (U[j1*dimX+i] - U[index]) + By[index]; + + val11 = fabs(val1) - denom_lam; if (val11 < 0) val11 = 0; + val22 = fabs(val2) - denom_lam; if (val22 < 0) val22 = 0; + + if (val1 !=0) Dx[index] = (val1/fabs(val1))*val11; else Dx[index] = 0; + if (val2 !=0) Dy[index] = (val2/fabs(val2))*val22; else Dy[index] = 0; + + }} + return 1; +} +float updDxDy_shrinkIso2D(float *U, float *Dx, float *Dy, float *Bx, float *By, long dimX, long dimY, float lambda) +{ + long i,j,i1,j1,index; + float val1, val11, val2, denom, denom_lam; + denom_lam = 1.0f/lambda; + +#pragma omp parallel for shared(U,denom_lam) private(index,i,j,i1,j1,val1,val11,val2,denom) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + index = j*dimX+i; + + val1 = (U[j*dimX+i1] - U[index]) + Bx[index]; + val2 = (U[j1*dimX+i] - U[index]) + By[index]; + + denom = sqrt(val1*val1 + val2*val2); + + val11 = (denom - denom_lam); if (val11 < 0) val11 = 0.0f; + + if (denom != 0.0f) { + Dx[index] = val11*(val1/denom); + Dy[index] = val11*(val2/denom); + } + else { + Dx[index] = 0; + Dy[index] = 0; + } + }} + return 1; +} +float updBxBy2D(float *U, float *Dx, float *Dy, float *Bx, float *By, long dimX, long dimY) +{ + long i,j,i1,j1,index; +#pragma omp parallel for shared(U) private(index,i,j,i1,j1) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + index = j*dimX+i; + + Bx[index] += (U[j*dimX+i1] - U[index]) - Dx[index]; + By[index] += (U[j1*dimX+i] - U[index]) - Dy[index]; + }} + return 1; +} + +/********************************************************************/ +/***************************3D Functions*****************************/ +/********************************************************************/ +/*****************************************************************/ +float gauss_seidel3D(float *U, float *A, float *U_prev, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, long dimX, long dimY, long dimZ, float lambda, float mu) +{ + float normConst, d_val, b_val, sum; + long i,j,i1,i2,j1,j2,k,k1,k2,index; + normConst = 1.0f/(mu + 6.0f*lambda); +#pragma omp parallel for shared(U) private(index,i,j,i1,i2,j1,j2,k,k1,k2,d_val,b_val,sum) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + k1 = k+1; if (k1 == dimZ) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + index = (dimX*dimY)*k + j*dimX+i; + + d_val = Dx[(dimX*dimY)*k + j*dimX+i2] - Dx[index] + Dy[(dimX*dimY)*k + j2*dimX+i] - Dy[index] + Dz[(dimX*dimY)*k2 + j*dimX+i] - Dz[index]; + b_val = -Bx[(dimX*dimY)*k + j*dimX+i2] + Bx[index] - By[(dimX*dimY)*k + j2*dimX+i] + By[index] - Bz[(dimX*dimY)*k2 + j*dimX+i] + Bz[index]; + sum = d_val + b_val; + sum += U_prev[(dimX*dimY)*k + j*dimX+i1] + U_prev[(dimX*dimY)*k + j*dimX+i2] + U_prev[(dimX*dimY)*k + j1*dimX+i] + U_prev[(dimX*dimY)*k + j2*dimX+i] + U_prev[(dimX*dimY)*k1 + j*dimX+i] + U_prev[(dimX*dimY)*k2 + j*dimX+i]; + sum *= lambda; + sum += mu*A[index]; + U[index] = normConst*sum; + }}} + return *U; +} + +float updDxDyDz_shrinkAniso3D(float *U, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, long dimX, long dimY, long dimZ, float lambda) +{ + long i,j,i1,j1,k,k1,index; + float val1, val11, val2, val22, val3, val33, denom_lam; + denom_lam = 1.0f/lambda; +#pragma omp parallel for shared(U,denom_lam) private(index,i,j,i1,j1,k,k1,val1,val11,val2,val22,val3,val33) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + k1 = k+1; if (k1 == dimZ) k1 = k-1; + + val1 = (U[(dimX*dimY)*k + j*dimX+i1] - U[index]) + Bx[index]; + val2 = (U[(dimX*dimY)*k + j1*dimX+i] - U[index]) + By[index]; + val3 = (U[(dimX*dimY)*k1 + j*dimX+i] - U[index]) + Bz[index]; + + val11 = fabs(val1) - denom_lam; if (val11 < 0.0f) val11 = 0.0f; + val22 = fabs(val2) - denom_lam; if (val22 < 0.0f) val22 = 0.0f; + val33 = fabs(val3) - denom_lam; if (val33 < 0.0f) val33 = 0.0f; + + if (val1 !=0.0f) Dx[index] = (val1/fabs(val1))*val11; else Dx[index] = 0.0f; + if (val2 !=0.0f) Dy[index] = (val2/fabs(val2))*val22; else Dy[index] = 0.0f; + if (val3 !=0.0f) Dz[index] = (val3/fabs(val3))*val33; else Dz[index] = 0.0f; + + }}} + return 1; +} +float updDxDyDz_shrinkIso3D(float *U, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, long dimX, long dimY, long dimZ, float lambda) +{ + long i,j,i1,j1,k,k1,index; + float val1, val11, val2, val3, denom, denom_lam; + denom_lam = 1.0f/lambda; +#pragma omp parallel for shared(U,denom_lam) private(index,denom,i,j,i1,j1,k,k1,val1,val11,val2,val3) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + k1 = k+1; if (k1 == dimZ) k1 = k-1; + + val1 = (U[(dimX*dimY)*k + j*dimX+i1] - U[index]) + Bx[index]; + val2 = (U[(dimX*dimY)*k + j1*dimX+i] - U[index]) + By[index]; + val3 = (U[(dimX*dimY)*k1 + j*dimX+i] - U[index]) + Bz[index]; + + denom = sqrt(val1*val1 + val2*val2 + val3*val3); + + val11 = (denom - denom_lam); if (val11 < 0) val11 = 0.0f; + + if (denom != 0.0f) { + Dx[index] = val11*(val1/denom); + Dy[index] = val11*(val2/denom); + Dz[index] = val11*(val3/denom); + } + else { + Dx[index] = 0; + Dy[index] = 0; + Dz[index] = 0; + } + }}} + return 1; +} +float updBxByBz3D(float *U, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, long dimX, long dimY, long dimZ) +{ + long i,j,k,i1,j1,k1,index; +#pragma omp parallel for shared(U) private(index,i,j,k,i1,j1,k1) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + k1 = k+1; if (k1 == dimZ) k1 = k-1; + + Bx[index] += (U[(dimX*dimY)*k + j*dimX+i1] - U[index]) - Dx[index]; + By[index] += (U[(dimX*dimY)*k + j1*dimX+i] - U[index]) - Dy[index]; + Bz[index] += (U[(dimX*dimY)*k1 + j*dimX+i] - U[index]) - Dz[index]; + }}} + return 1; +} diff --git a/src/Core/regularisers_CPU/SB_TV_core.h b/src/Core/regularisers_CPU/SB_TV_core.h new file mode 100644 index 0000000..7485e3b --- /dev/null +++ b/src/Core/regularisers_CPU/SB_TV_core.h @@ -0,0 +1,61 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + + +/* C-OMP implementation of Split Bregman - TV denoising-regularisation model (2D/3D) [1] +* +* Input Parameters: +* 1. Noisy image/volume +* 2. lambda - regularisation parameter +* 3. Number of iterations [OPTIONAL parameter] +* 4. eplsilon - tolerance constant [OPTIONAL parameter] +* 5. TV-type: 'iso' or 'l1' [OPTIONAL parameter] +* 6. print information: 0 (off) or 1 (on) [OPTIONAL parameter] +* +* Output: +* 1. Filtered/regularized image +* +* [1]. Goldstein, T. and Osher, S., 2009. The split Bregman method for L1-regularized problems. SIAM journal on imaging sciences, 2(2), pp.323-343. +*/ + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float SB_TV_CPU_main(float *Input, float *Output, float mu, int iter, float epsil, int methodTV, int printM, int dimX, int dimY, int dimZ); + +CCPI_EXPORT float gauss_seidel2D(float *U, float *A, float *U_prev, float *Dx, float *Dy, float *Bx, float *By, long dimX, long dimY, float lambda, float mu); +CCPI_EXPORT float updDxDy_shrinkAniso2D(float *U, float *Dx, float *Dy, float *Bx, float *By, long dimX, long dimY, float lambda); +CCPI_EXPORT float updDxDy_shrinkIso2D(float *U, float *Dx, float *Dy, float *Bx, float *By, long dimX, long dimY, float lambda); +CCPI_EXPORT float updBxBy2D(float *U, float *Dx, float *Dy, float *Bx, float *By, long dimX, long dimY); + +CCPI_EXPORT float gauss_seidel3D(float *U, float *A, float *U_prev, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, long dimX, long dimY, long dimZ, float lambda, float mu); +CCPI_EXPORT float updDxDyDz_shrinkAniso3D(float *U, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, long dimX, long dimY, long dimZ, float lambda); +CCPI_EXPORT float updDxDyDz_shrinkIso3D(float *U, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, long dimX, long dimY, long dimZ, float lambda); +CCPI_EXPORT float updBxByBz3D(float *U, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, long dimX, long dimY, long dimZ); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/TGV_core.c b/src/Core/regularisers_CPU/TGV_core.c new file mode 100644 index 0000000..805c3d4 --- /dev/null +++ b/src/Core/regularisers_CPU/TGV_core.c @@ -0,0 +1,487 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "TGV_core.h" + +/* C-OMP implementation of Primal-Dual denoising method for + * Total Generilized Variation (TGV)-L2 model [1] (2D/3D case) + * + * Input Parameters: + * 1. Noisy image/volume (2D/3D) + * 2. lambda - regularisation parameter + * 3. parameter to control the first-order term (alpha1) + * 4. parameter to control the second-order term (alpha0) + * 5. Number of Chambolle-Pock (Primal-Dual) iterations + * 6. Lipshitz constant (default is 12) + * + * Output: + * Filtered/regularised image/volume + * + * References: + * [1] K. Bredies "Total Generalized Variation" + * + */ + +float TGV_main(float *U0, float *U, float lambda, float alpha1, float alpha0, int iter, float L2, int dimX, int dimY, int dimZ) +{ + long DimTotal; + int ll; + float *U_old, *P1, *P2, *Q1, *Q2, *Q3, *V1, *V1_old, *V2, *V2_old, tau, sigma; + + DimTotal = (long)(dimX*dimY*dimZ); + copyIm(U0, U, (long)(dimX), (long)(dimY), (long)(dimZ)); /* initialize */ + tau = pow(L2,-0.5); + sigma = pow(L2,-0.5); + + /* dual variables */ + P1 = calloc(DimTotal, sizeof(float)); + P2 = calloc(DimTotal, sizeof(float)); + + Q1 = calloc(DimTotal, sizeof(float)); + Q2 = calloc(DimTotal, sizeof(float)); + Q3 = calloc(DimTotal, sizeof(float)); + + U_old = calloc(DimTotal, sizeof(float)); + + V1 = calloc(DimTotal, sizeof(float)); + V1_old = calloc(DimTotal, sizeof(float)); + V2 = calloc(DimTotal, sizeof(float)); + V2_old = calloc(DimTotal, sizeof(float)); + + if (dimZ == 1) { + /*2D case*/ + + /* Primal-dual iterations begin here */ + for(ll = 0; ll < iter; ll++) { + + /* Calculate Dual Variable P */ + DualP_2D(U, V1, V2, P1, P2, (long)(dimX), (long)(dimY), sigma); + + /*Projection onto convex set for P*/ + ProjP_2D(P1, P2, (long)(dimX), (long)(dimY), alpha1); + + /* Calculate Dual Variable Q */ + DualQ_2D(V1, V2, Q1, Q2, Q3, (long)(dimX), (long)(dimY), sigma); + + /*Projection onto convex set for Q*/ + ProjQ_2D(Q1, Q2, Q3, (long)(dimX), (long)(dimY), alpha0); + + /*saving U into U_old*/ + copyIm(U, U_old, (long)(dimX), (long)(dimY), 1l); + + /*adjoint operation -> divergence and projection of P*/ + DivProjP_2D(U, U0, P1, P2, (long)(dimX), (long)(dimY), lambda, tau); + + /*get updated solution U*/ + newU(U, U_old, (long)(dimX), (long)(dimY)); + + /*saving V into V_old*/ + copyIm(V1, V1_old, (long)(dimX), (long)(dimY), 1l); + copyIm(V2, V2_old, (long)(dimX), (long)(dimY), 1l); + + /* upd V*/ + UpdV_2D(V1, V2, P1, P2, Q1, Q2, Q3, (long)(dimX), (long)(dimY), tau); + + /*get new V*/ + newU(V1, V1_old, (long)(dimX), (long)(dimY)); + newU(V2, V2_old, (long)(dimX), (long)(dimY)); + } /*end of iterations*/ + } + else { + /*3D case*/ + float *P3, *Q4, *Q5, *Q6, *V3, *V3_old; + + P3 = calloc(DimTotal, sizeof(float)); + Q4 = calloc(DimTotal, sizeof(float)); + Q5 = calloc(DimTotal, sizeof(float)); + Q6 = calloc(DimTotal, sizeof(float)); + V3 = calloc(DimTotal, sizeof(float)); + V3_old = calloc(DimTotal, sizeof(float)); + + /* Primal-dual iterations begin here */ + for(ll = 0; ll < iter; ll++) { + + /* Calculate Dual Variable P */ + DualP_3D(U, V1, V2, V3, P1, P2, P3, (long)(dimX), (long)(dimY), (long)(dimZ), sigma); + + /*Projection onto convex set for P*/ + ProjP_3D(P1, P2, P3, (long)(dimX), (long)(dimY), (long)(dimZ), alpha1); + + /* Calculate Dual Variable Q */ + DualQ_3D(V1, V2, V3, Q1, Q2, Q3, Q4, Q5, Q6, (long)(dimX), (long)(dimY), (long)(dimZ), sigma); + + /*Projection onto convex set for Q*/ + ProjQ_3D(Q1, Q2, Q3, Q4, Q5, Q6, (long)(dimX), (long)(dimY), (long)(dimZ), alpha0); + + /*saving U into U_old*/ + copyIm(U, U_old, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /*adjoint operation -> divergence and projection of P*/ + DivProjP_3D(U, U0, P1, P2, P3, (long)(dimX), (long)(dimY), (long)(dimZ), lambda, tau); + + /*get updated solution U*/ + newU3D(U, U_old, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /*saving V into V_old*/ + copyIm_3Ar(V1, V2, V3, V1_old, V2_old, V3_old, (long)(dimX), (long)(dimY), (long)(dimZ)); + + /* upd V*/ + UpdV_3D(V1, V2, V3, P1, P2, P3, Q1, Q2, Q3, Q4, Q5, Q6, (long)(dimX), (long)(dimY), (long)(dimZ), tau); + + /*get new V*/ + newU3D_3Ar(V1, V2, V3, V1_old, V2_old, V3_old, (long)(dimX), (long)(dimY), (long)(dimZ)); + } /*end of iterations*/ + free(P3);free(Q4);free(Q5);free(Q6);free(V3);free(V3_old); + } + + /*freeing*/ + free(P1);free(P2);free(Q1);free(Q2);free(Q3);free(U_old); + free(V1);free(V2);free(V1_old);free(V2_old); + return *U; +} + +/********************************************************************/ +/***************************2D Functions*****************************/ +/********************************************************************/ + +/*Calculating dual variable P (using forward differences)*/ +float DualP_2D(float *U, float *V1, float *V2, float *P1, float *P2, long dimX, long dimY, float sigma) +{ + long i,j, index; +#pragma omp parallel for shared(U,V1,V2,P1,P2) private(i,j,index) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + if (i == dimX-1) P1[index] += sigma*((U[j*dimX+(i-1)] - U[index]) - V1[index]); + else P1[index] += sigma*((U[j*dimX+(i+1)] - U[index]) - V1[index]); + if (j == dimY-1) P2[index] += sigma*((U[(j-1)*dimX+i] - U[index]) - V2[index]); + else P2[index] += sigma*((U[(j+1)*dimX+i] - U[index]) - V2[index]); + }} + return 1; +} +/*Projection onto convex set for P*/ +float ProjP_2D(float *P1, float *P2, long dimX, long dimY, float alpha1) +{ + float grad_magn; + long i,j,index; +#pragma omp parallel for shared(P1,P2) private(i,j,index,grad_magn) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + grad_magn = (sqrtf(pow(P1[index],2) + pow(P2[index],2)))/alpha1; + if (grad_magn > 1.0f) { + P1[index] /= grad_magn; + P2[index] /= grad_magn; + } + }} + return 1; +} +/*Calculating dual variable Q (using forward differences)*/ +float DualQ_2D(float *V1, float *V2, float *Q1, float *Q2, float *Q3, long dimX, long dimY, float sigma) +{ + long i,j,index; + float q1, q2, q11, q22; +#pragma omp parallel for shared(Q1,Q2,Q3,V1,V2) private(i,j,index,q1,q2,q11,q22) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + q1 = 0.0f; q11 = 0.0f; q2 = 0.0f; q22 = 0.0f; + /* boundary conditions (Neuman) */ + if (i != dimX-1){ + q1 = V1[j*dimX+(i+1)] - V1[index]; + q11 = V2[j*dimX+(i+1)] - V2[index]; + } + if (j != dimY-1) { + q2 = V2[(j+1)*dimX+i] - V2[index]; + q22 = V1[(j+1)*dimX+i] - V1[index]; + } + Q1[index] += sigma*(q1); + Q2[index] += sigma*(q2); + Q3[index] += sigma*(0.5f*(q11 + q22)); + }} + return 1; +} +float ProjQ_2D(float *Q1, float *Q2, float *Q3, long dimX, long dimY, float alpha0) +{ + float grad_magn; + long i,j,index; +#pragma omp parallel for shared(Q1,Q2,Q3) private(i,j,index,grad_magn) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + grad_magn = sqrtf(pow(Q1[index],2) + pow(Q2[index],2) + 2*pow(Q3[index],2)); + grad_magn = grad_magn/alpha0; + if (grad_magn > 1.0f) { + Q1[index] /= grad_magn; + Q2[index] /= grad_magn; + Q3[index] /= grad_magn; + } + }} + return 1; +} +/* Divergence and projection for P*/ +float DivProjP_2D(float *U, float *U0, float *P1, float *P2, long dimX, long dimY, float lambda, float tau) +{ + long i,j,index; + float P_v1, P_v2, div; +#pragma omp parallel for shared(U,U0,P1,P2) private(i,j,index,P_v1,P_v2,div) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + if (i == 0) P_v1 = P1[index]; + else P_v1 = P1[index] - P1[j*dimX+(i-1)]; + if (j == 0) P_v2 = P2[index]; + else P_v2 = P2[index] - P2[(j-1)*dimX+i]; + div = P_v1 + P_v2; + U[index] = (lambda*(U[index] + tau*div) + tau*U0[index])/(lambda + tau); + }} + return *U; +} +/*get updated solution U*/ +float newU(float *U, float *U_old, long dimX, long dimY) +{ + long i; +#pragma omp parallel for shared(U,U_old) private(i) + for(i=0; i<dimX*dimY; i++) U[i] = 2*U[i] - U_old[i]; + return *U; +} +/*get update for V*/ +float UpdV_2D(float *V1, float *V2, float *P1, float *P2, float *Q1, float *Q2, float *Q3, long dimX, long dimY, float tau) +{ + long i, j, index; + float q1, q3_x, q3_y, q2, div1, div2; +#pragma omp parallel for shared(V1,V2,P1,P2,Q1,Q2,Q3) private(i, j, index, q1, q3_x, q3_y, q2, div1, div2) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + index = j*dimX+i; + q2 = 0.0f; q3_y = 0.0f; q1 = 0.0f; q3_x = 0.0; + /* boundary conditions (Neuman) */ + if (i != 0) { + q1 = Q1[index] - Q1[j*dimX+(i-1)]; + q3_x = Q3[index] - Q3[j*dimX+(i-1)]; + } + if (j != 0) { + q2 = Q2[index] - Q2[(j-1)*dimX+i]; + q3_y = Q3[index] - Q3[(j-1)*dimX+i]; + } + div1 = q1 + q3_y; + div2 = q3_x + q2; + V1[index] += tau*(P1[index] + div1); + V2[index] += tau*(P2[index] + div2); + }} + return 1; +} + +/********************************************************************/ +/***************************3D Functions*****************************/ +/********************************************************************/ +/*Calculating dual variable P (using forward differences)*/ +float DualP_3D(float *U, float *V1, float *V2, float *V3, float *P1, float *P2, float *P3, long dimX, long dimY, long dimZ, float sigma) +{ + long i,j,k, index; +#pragma omp parallel for shared(U,V1,V2,V3,P1,P2,P3) private(i,j,k,index) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + if (i == dimX-1) P1[index] += sigma*((U[(dimX*dimY)*k + j*dimX+(i-1)] - U[index]) - V1[index]); + else P1[index] += sigma*((U[(dimX*dimY)*k + j*dimX+(i+1)] - U[index]) - V1[index]); + if (j == dimY-1) P2[index] += sigma*((U[(dimX*dimY)*k + (j-1)*dimX+i] - U[index]) - V2[index]); + else P2[index] += sigma*((U[(dimX*dimY)*k + (j+1)*dimX+i] - U[index]) - V2[index]); + if (k == dimZ-1) P3[index] += sigma*((U[(dimX*dimY)*(k-1) + j*dimX+i] - U[index]) - V3[index]); + else P3[index] += sigma*((U[(dimX*dimY)*(k+1) + j*dimX+i] - U[index]) - V3[index]); + }}} + return 1; +} +/*Projection onto convex set for P*/ +float ProjP_3D(float *P1, float *P2, float *P3, long dimX, long dimY, long dimZ, float alpha1) +{ + float grad_magn; + long i,j,k,index; +#pragma omp parallel for shared(P1,P2,P3) private(i,j,k,index,grad_magn) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + grad_magn = (sqrtf(pow(P1[index],2) + pow(P2[index],2) + pow(P3[index],2)))/alpha1; + if (grad_magn > 1.0f) { + P1[index] /= grad_magn; + P2[index] /= grad_magn; + P3[index] /= grad_magn; + } + }}} + return 1; +} +/*Calculating dual variable Q (using forward differences)*/ +float DualQ_3D(float *V1, float *V2, float *V3, float *Q1, float *Q2, float *Q3, float *Q4, float *Q5, float *Q6, long dimX, long dimY, long dimZ, float sigma) +{ + long i,j,k,index; + float q1, q2, q3, q11, q22, q33, q44, q55, q66; +#pragma omp parallel for shared(Q1,Q2,Q3,Q4,Q5,Q6,V1,V2,V3) private(i,j,k,index,q1,q2,q3,q11,q22,q33,q44,q55,q66) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + q1 = 0.0f; q11 = 0.0f; q33 = 0.0f; q2 = 0.0f; q22 = 0.0f; q55 = 0.0f; q3 = 0.0f; q44 = 0.0f; q66 = 0.0f; + /* symmetric boundary conditions (Neuman) */ + if (i != dimX-1){ + q1 = V1[(dimX*dimY)*k + j*dimX+(i+1)] - V1[index]; + q11 = V2[(dimX*dimY)*k + j*dimX+(i+1)] - V2[index]; + q33 = V3[(dimX*dimY)*k + j*dimX+(i+1)] - V3[index]; + } + if (j != dimY-1) { + q2 = V2[(dimX*dimY)*k + (j+1)*dimX+i] - V2[index]; + q22 = V1[(dimX*dimY)*k + (j+1)*dimX+i] - V1[index]; + q55 = V3[(dimX*dimY)*k + (j+1)*dimX+i] - V3[index]; + } + if (k != dimZ-1) { + q3 = V3[(dimX*dimY)*(k+1) + j*dimX+i] - V3[index]; + q44 = V1[(dimX*dimY)*(k+1) + j*dimX+i] - V1[index]; + q66 = V2[(dimX*dimY)*(k+1) + j*dimX+i] - V2[index]; + } + + Q1[index] += sigma*(q1); /*Q11*/ + Q2[index] += sigma*(q2); /*Q22*/ + Q3[index] += sigma*(q3); /*Q33*/ + Q4[index] += sigma*(0.5f*(q11 + q22)); /* Q21 / Q12 */ + Q5[index] += sigma*(0.5f*(q33 + q44)); /* Q31 / Q13 */ + Q6[index] += sigma*(0.5f*(q55 + q66)); /* Q32 / Q23 */ + }}} + return 1; +} +float ProjQ_3D(float *Q1, float *Q2, float *Q3, float *Q4, float *Q5, float *Q6, long dimX, long dimY, long dimZ, float alpha0) +{ + float grad_magn; + long i,j,k,index; +#pragma omp parallel for shared(Q1,Q2,Q3,Q4,Q5,Q6) private(i,j,k,index,grad_magn) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + grad_magn = sqrtf(pow(Q1[index],2) + pow(Q2[index],2) + pow(Q3[index],2) + 2.0f*pow(Q4[index],2) + 2.0f*pow(Q5[index],2) + 2.0f*pow(Q6[index],2)); + grad_magn = grad_magn/alpha0; + if (grad_magn > 1.0f) { + Q1[index] /= grad_magn; + Q2[index] /= grad_magn; + Q3[index] /= grad_magn; + Q4[index] /= grad_magn; + Q5[index] /= grad_magn; + Q6[index] /= grad_magn; + } + }}} + return 1; +} +/* Divergence and projection for P*/ +float DivProjP_3D(float *U, float *U0, float *P1, float *P2, float *P3, long dimX, long dimY, long dimZ, float lambda, float tau) +{ + long i,j,k,index; + float P_v1, P_v2, P_v3, div; +#pragma omp parallel for shared(U,U0,P1,P2,P3) private(i,j,k,index,P_v1,P_v2,P_v3,div) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + if (i == 0) P_v1 = P1[index]; + else P_v1 = P1[index] - P1[(dimX*dimY)*k + j*dimX+(i-1)]; + if (j == 0) P_v2 = P2[index]; + else P_v2 = P2[index] - P2[(dimX*dimY)*k + (j-1)*dimX+i]; + if (k == 0) P_v3 = P3[index]; + else P_v3 = P3[index] - P3[(dimX*dimY)*(k-1) + (j)*dimX+i]; + + div = P_v1 + P_v2 + P_v3; + U[index] = (lambda*(U[index] + tau*div) + tau*U0[index])/(lambda + tau); + }}} + return *U; +} +/*get update for V*/ +float UpdV_3D(float *V1, float *V2, float *V3, float *P1, float *P2, float *P3, float *Q1, float *Q2, float *Q3, float *Q4, float *Q5, float *Q6, long dimX, long dimY, long dimZ, float tau) +{ + long i,j,k,index; + float q1, q4x, q5x, q2, q4y, q6y, q6z, q5z, q3, div1, div2, div3; +#pragma omp parallel for shared(V1,V2,V3,P1,P2,P3,Q1,Q2,Q3,Q4,Q5,Q6) private(i,j,k,index,q1,q4x,q5x,q2,q4y,q6y,q6z,q5z,q3,div1,div2,div3) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + for(k=0; k<dimZ; k++) { + index = (dimX*dimY)*k + j*dimX+i; + q1 = 0.0f; q4x= 0.0f; q5x= 0.0f; q2= 0.0f; q4y= 0.0f; q6y= 0.0f; q6z= 0.0f; q5z= 0.0f; q3= 0.0f; + /* Q1 - Q11, Q2 - Q22, Q3 - Q33, Q4 - Q21/Q12, Q5 - Q31/Q13, Q6 - Q32/Q23*/ + /* symmetric boundary conditions (Neuman) */ + if (i != 0) { + q1 = Q1[index] - Q1[(dimX*dimY)*k + j*dimX+(i-1)]; + q4x = Q4[index] - Q4[(dimX*dimY)*k + j*dimX+(i-1)]; + q5x = Q5[index] - Q5[(dimX*dimY)*k + j*dimX+(i-1)]; + } + if (j != 0) { + q2 = Q2[index] - Q2[(dimX*dimY)*k + (j-1)*dimX+i]; + q4y = Q4[index] - Q4[(dimX*dimY)*k + (j-1)*dimX+i]; + q6y = Q6[index] - Q6[(dimX*dimY)*k + (j-1)*dimX+i]; + } + if (k != 0) { + q6z = Q6[index] - Q6[(dimX*dimY)*(k-1) + (j)*dimX+i]; + q5z = Q5[index] - Q5[(dimX*dimY)*(k-1) + (j)*dimX+i]; + q3 = Q3[index] - Q3[(dimX*dimY)*(k-1) + (j)*dimX+i]; + } + div1 = q1 + q4y + q5z; + div2 = q4x + q2 + q6z; + div3 = q5x + q6y + q3; + + V1[index] += tau*(P1[index] + div1); + V2[index] += tau*(P2[index] + div2); + V3[index] += tau*(P3[index] + div3); + }}} + return 1; +} + +float copyIm_3Ar(float *V1, float *V2, float *V3, float *V1_old, float *V2_old, float *V3_old, long dimX, long dimY, long dimZ) +{ + long j; +#pragma omp parallel for shared(V1, V2, V3, V1_old, V2_old, V3_old) private(j) + for (j = 0; j<dimX*dimY*dimZ; j++) { + V1_old[j] = V1[j]; + V2_old[j] = V2[j]; + V3_old[j] = V3[j]; + } + return 1; +} + +/*get updated solution U*/ +float newU3D(float *U, float *U_old, long dimX, long dimY, long dimZ) +{ + long i; +#pragma omp parallel for shared(U, U_old) private(i) + for(i=0; i<dimX*dimY*dimZ; i++) U[i] = 2.0f*U[i] - U_old[i]; + return *U; +} + + +/*get updated solution U*/ +float newU3D_3Ar(float *V1, float *V2, float *V3, float *V1_old, float *V2_old, float *V3_old, long dimX, long dimY, long dimZ) +{ + long i; +#pragma omp parallel for shared(V1, V2, V3, V1_old, V2_old, V3_old) private(i) + for(i=0; i<dimX*dimY*dimZ; i++) { + V1[i] = 2.0f*V1[i] - V1_old[i]; + V2[i] = 2.0f*V2[i] - V2_old[i]; + V3[i] = 2.0f*V3[i] - V3_old[i]; + } + return 1; +} + diff --git a/src/Core/regularisers_CPU/TGV_core.h b/src/Core/regularisers_CPU/TGV_core.h new file mode 100644 index 0000000..11b12c1 --- /dev/null +++ b/src/Core/regularisers_CPU/TGV_core.h @@ -0,0 +1,73 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + +/* C-OMP implementation of Primal-Dual denoising method for + * Total Generilized Variation (TGV)-L2 model [1] (2D/3D) + * + * Input Parameters: + * 1. Noisy image/volume (2D/3D) + * 2. lambda - regularisation parameter + * 3. parameter to control the first-order term (alpha1) + * 4. parameter to control the second-order term (alpha0) + * 5. Number of Chambolle-Pock (Primal-Dual) iterations + * 6. Lipshitz constant (default is 12) + * + * Output: + * Filtered/regularised image/volume + * + * References: + * [1] K. Bredies "Total Generalized Variation" + */ + + +#ifdef __cplusplus +extern "C" { +#endif + +CCPI_EXPORT float TGV_main(float *U0, float *U, float lambda, float alpha1, float alpha0, int iter, float L2, int dimX, int dimY, int dimZ); + +/* 2D functions */ +CCPI_EXPORT float DualP_2D(float *U, float *V1, float *V2, float *P1, float *P2, long dimX, long dimY, float sigma); +CCPI_EXPORT float ProjP_2D(float *P1, float *P2, long dimX, long dimY, float alpha1); +CCPI_EXPORT float DualQ_2D(float *V1, float *V2, float *Q1, float *Q2, float *Q3, long dimX, long dimY, float sigma); +CCPI_EXPORT float ProjQ_2D(float *Q1, float *Q2, float *Q3, long dimX, long dimY, float alpha0); +CCPI_EXPORT float DivProjP_2D(float *U, float *U0, float *P1, float *P2, long dimX, long dimY, float lambda, float tau); +CCPI_EXPORT float UpdV_2D(float *V1, float *V2, float *P1, float *P2, float *Q1, float *Q2, float *Q3, long dimX, long dimY, float tau); +CCPI_EXPORT float newU(float *U, float *U_old, long dimX, long dimY); +/* 3D functions */ +CCPI_EXPORT float DualP_3D(float *U, float *V1, float *V2, float *V3, float *P1, float *P2, float *P3, long dimX, long dimY, long dimZ, float sigma); +CCPI_EXPORT float ProjP_3D(float *P1, float *P2, float *P3, long dimX, long dimY, long dimZ, float alpha1); +CCPI_EXPORT float DualQ_3D(float *V1, float *V2, float *V3, float *Q1, float *Q2, float *Q3, float *Q4, float *Q5, float *Q6, long dimX, long dimY, long dimZ, float sigma); +CCPI_EXPORT float ProjQ_3D(float *Q1, float *Q2, float *Q3, float *Q4, float *Q5, float *Q6, long dimX, long dimY, long dimZ, float alpha0); +CCPI_EXPORT float DivProjP_3D(float *U, float *U0, float *P1, float *P2, float *P3, long dimX, long dimY, long dimZ, float lambda, float tau); +CCPI_EXPORT float UpdV_3D(float *V1, float *V2, float *V3, float *P1, float *P2, float *P3, float *Q1, float *Q2, float *Q3, float *Q4, float *Q5, float *Q6, long dimX, long dimY, long dimZ, float tau); +CCPI_EXPORT float newU3D(float *U, float *U_old, long dimX, long dimY, long dimZ); +CCPI_EXPORT float copyIm_3Ar(float *V1, float *V2, float *V3, float *V1_old, float *V2_old, float *V3_old, long dimX, long dimY, long dimZ); +CCPI_EXPORT float newU3D_3Ar(float *V1, float *V2, float *V3, float *V1_old, float *V2_old, float *V3_old, long dimX, long dimY, long dimZ); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_CPU/TNV_core.c b/src/Core/regularisers_CPU/TNV_core.c new file mode 100755 index 0000000..753cc5f --- /dev/null +++ b/src/Core/regularisers_CPU/TNV_core.c @@ -0,0 +1,452 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "TNV_core.h" + +/* + * C-OMP implementation of Total Nuclear Variation regularisation model (2D + channels) [1] + * The code is modified from the implementation by Joan Duran <joan.duran@uib.es> see + * "denoisingPDHG_ipol.cpp" in Joans Collaborative Total Variation package + * + * Input Parameters: + * 1. Noisy volume of 2D + channel dimension, i.e. 3D volume + * 2. lambda - regularisation parameter + * 3. Number of iterations [OPTIONAL parameter] + * 4. eplsilon - tolerance constant [OPTIONAL parameter] + * 5. print information: 0 (off) or 1 (on) [OPTIONAL parameter] + * + * Output: + * 1. Filtered/regularized image + * + * [1]. Duran, J., Moeller, M., Sbert, C. and Cremers, D., 2016. Collaborative total variation: a general framework for vectorial TV models. SIAM Journal on Imaging Sciences, 9(1), pp.116-151. + */ + +float TNV_CPU_main(float *Input, float *u, float lambda, int maxIter, float tol, int dimX, int dimY, int dimZ) +{ + long k, p, q, r, DimTotal; + float taulambda; + float *u_upd, *gx, *gy, *gx_upd, *gy_upd, *qx, *qy, *qx_upd, *qy_upd, *v, *vx, *vy, *gradx, *grady, *gradx_upd, *grady_upd, *gradx_ubar, *grady_ubar, *div, *div_upd; + + p = 1l; + q = 1l; + r = 0l; + + lambda = 1.0f/(2.0f*lambda); + DimTotal = (long)(dimX*dimY*dimZ); + /* PDHG algorithm parameters*/ + float tau = 0.5f; + float sigma = 0.5f; + float theta = 1.0f; + + // Auxiliar vectors + u_upd = calloc(DimTotal, sizeof(float)); + gx = calloc(DimTotal, sizeof(float)); + gy = calloc(DimTotal, sizeof(float)); + gx_upd = calloc(DimTotal, sizeof(float)); + gy_upd = calloc(DimTotal, sizeof(float)); + qx = calloc(DimTotal, sizeof(float)); + qy = calloc(DimTotal, sizeof(float)); + qx_upd = calloc(DimTotal, sizeof(float)); + qy_upd = calloc(DimTotal, sizeof(float)); + v = calloc(DimTotal, sizeof(float)); + vx = calloc(DimTotal, sizeof(float)); + vy = calloc(DimTotal, sizeof(float)); + gradx = calloc(DimTotal, sizeof(float)); + grady = calloc(DimTotal, sizeof(float)); + gradx_upd = calloc(DimTotal, sizeof(float)); + grady_upd = calloc(DimTotal, sizeof(float)); + gradx_ubar = calloc(DimTotal, sizeof(float)); + grady_ubar = calloc(DimTotal, sizeof(float)); + div = calloc(DimTotal, sizeof(float)); + div_upd = calloc(DimTotal, sizeof(float)); + + // Backtracking parameters + float s = 1.0f; + float gamma = 0.75f; + float beta = 0.95f; + float alpha0 = 0.2f; + float alpha = alpha0; + float delta = 1.5f; + float eta = 0.95f; + + // PDHG algorithm parameters + taulambda = tau * lambda; + float divtau = 1.0f / tau; + float divsigma = 1.0f / sigma; + float theta1 = 1.0f + theta; + + /*allocate memory for taulambda */ + //taulambda = (float*) calloc(dimZ, sizeof(float)); + //for(k=0; k < dimZ; k++) {taulambda[k] = tau*lambda[k];} + + // Apply Primal-Dual Hybrid Gradient scheme + int iter = 0; + float residual = fLarge; + float ubarx, ubary; + + for(iter = 0; iter < maxIter; iter++) { + // Argument of proximal mapping of fidelity term +#pragma omp parallel for shared(v, u) private(k) + for(k=0; k<dimX*dimY*dimZ; k++) {v[k] = u[k] + tau*div[k];} + +// Proximal solution of fidelity term +proxG(u_upd, v, Input, taulambda, (long)(dimX), (long)(dimY), (long)(dimZ)); + +// Gradient of updated primal variable +gradient(u_upd, gradx_upd, grady_upd, (long)(dimX), (long)(dimY), (long)(dimZ)); + +// Argument of proximal mapping of regularization term +#pragma omp parallel for shared(gradx_upd, grady_upd, gradx, grady) private(k, ubarx, ubary) +for(k=0; k<dimX*dimY*dimZ; k++) { + ubarx = theta1 * gradx_upd[k] - theta * gradx[k]; + ubary = theta1 * grady_upd[k] - theta * grady[k]; + vx[k] = ubarx + divsigma * qx[k]; + vy[k] = ubary + divsigma * qy[k]; + gradx_ubar[k] = ubarx; + grady_ubar[k] = ubary; +} + +proxF(gx_upd, gy_upd, vx, vy, sigma, p, q, r, (long)(dimX), (long)(dimY), (long)(dimZ)); + +// Update dual variable +#pragma omp parallel for shared(qx_upd, qy_upd) private(k) +for(k=0; k<dimX*dimY*dimZ; k++) { + qx_upd[k] = qx[k] + sigma * (gradx_ubar[k] - gx_upd[k]); + qy_upd[k] = qy[k] + sigma * (grady_ubar[k] - gy_upd[k]); +} + +// Divergence of updated dual variable +#pragma omp parallel for shared(div_upd) private(k) +for(k=0; k<dimX*dimY*dimZ; k++) {div_upd[k] = 0.0f;} +divergence(qx_upd, qy_upd, div_upd, dimX, dimY, dimZ); + +// Compute primal residual, dual residual, and backtracking condition +float resprimal = 0.0f; +float resdual = 0.0f; +float product = 0.0f; +float unorm = 0.0f; +float qnorm = 0.0f; + +for(k=0; k<dimX*dimY*dimZ; k++) { + float udiff = u[k] - u_upd[k]; + float qxdiff = qx[k] - qx_upd[k]; + float qydiff = qy[k] - qy_upd[k]; + float divdiff = div[k] - div_upd[k]; + float gradxdiff = gradx[k] - gradx_upd[k]; + float gradydiff = grady[k] - grady_upd[k]; + + resprimal += fabs(divtau*udiff + divdiff); + resdual += fabs(divsigma*qxdiff - gradxdiff); + resdual += fabs(divsigma*qydiff - gradydiff); + + unorm += (udiff * udiff); + qnorm += (qxdiff * qxdiff + qydiff * qydiff); + product += (gradxdiff * qxdiff + gradydiff * qydiff); +} + +float b = (2.0f * tau * sigma * product) / (gamma * sigma * unorm + + gamma * tau * qnorm); + +// Adapt step-size parameters +float dual_dot_delta = resdual * s * delta; +float dual_div_delta = (resdual * s) / delta; + +if(b > 1) +{ + // Decrease step-sizes to fit balancing principle + tau = (beta * tau) / b; + sigma = (beta * sigma) / b; + alpha = alpha0; + + copyIm(u, u_upd, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(gx, gx_upd, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(gy, gy_upd, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(qx, qx_upd, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(qy, qy_upd, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(gradx, gradx_upd, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(grady, grady_upd, (long)(dimX), (long)(dimY), (long)(dimZ)); + copyIm(div, div_upd, (long)(dimX), (long)(dimY), (long)(dimZ)); + +} else if(resprimal > dual_dot_delta) +{ + // Increase primal step-size and decrease dual step-size + tau = tau / (1.0f - alpha); + sigma = sigma * (1.0f - alpha); + alpha = alpha * eta; + +} else if(resprimal < dual_div_delta) +{ + // Decrease primal step-size and increase dual step-size + tau = tau * (1.0f - alpha); + sigma = sigma / (1.0f - alpha); + alpha = alpha * eta; +} + +// Update variables +taulambda = tau * lambda; +//for(k=0; k < dimZ; k++) taulambda[k] = tau*lambda[k]; + +divsigma = 1.0f / sigma; +divtau = 1.0f / tau; + +copyIm(u_upd, u, (long)(dimX), (long)(dimY), (long)(dimZ)); +copyIm(gx_upd, gx, (long)(dimX), (long)(dimY), (long)(dimZ)); +copyIm(gy_upd, gy, (long)(dimX), (long)(dimY), (long)(dimZ)); +copyIm(qx_upd, qx, (long)(dimX), (long)(dimY), (long)(dimZ)); +copyIm(qy_upd, qy, (long)(dimX), (long)(dimY), (long)(dimZ)); +copyIm(gradx_upd, gradx, (long)(dimX), (long)(dimY), (long)(dimZ)); +copyIm(grady_upd, grady, (long)(dimX), (long)(dimY), (long)(dimZ)); +copyIm(div_upd, div, (long)(dimX), (long)(dimY), (long)(dimZ)); + +// Compute residual at current iteration +residual = (resprimal + resdual) / ((float) (dimX*dimY*dimZ)); + +// printf("%f \n", residual); +if (residual < tol) { + printf("Iterations stopped at %i with the residual %f \n", iter, residual); + break; } + + } + printf("Iterations stopped at %i with the residual %f \n", iter, residual); + free (u_upd); free(gx); free(gy); free(gx_upd); free(gy_upd); + free(qx); free(qy); free(qx_upd); free(qy_upd); free(v); free(vx); free(vy); + free(gradx); free(grady); free(gradx_upd); free(grady_upd); free(gradx_ubar); + free(grady_ubar); free(div); free(div_upd); + return *u; +} + +float proxG(float *u_upd, float *v, float *f, float taulambda, long dimX, long dimY, long dimZ) +{ + float constant; + long k; + constant = 1.0f + taulambda; +#pragma omp parallel for shared(v, f, u_upd) private(k) + for(k=0; k<dimZ*dimX*dimY; k++) { + u_upd[k] = (v[k] + taulambda * f[k])/constant; + //u_upd[(dimX*dimY)*k + l] = (v[(dimX*dimY)*k + l] + taulambda * f[(dimX*dimY)*k + l])/constant; + } + return *u_upd; +} + +float gradient(float *u_upd, float *gradx_upd, float *grady_upd, long dimX, long dimY, long dimZ) +{ + long i, j, k, l; + // Compute discrete gradient using forward differences +#pragma omp parallel for shared(gradx_upd,grady_upd,u_upd) private(i, j, k, l) + for(k = 0; k < dimZ; k++) { + for(j = 0; j < dimY; j++) { + l = j * dimX; + for(i = 0; i < dimX; i++) { + // Derivatives in the x-direction + if(i != dimX-1) + gradx_upd[(dimX*dimY)*k + i+l] = u_upd[(dimX*dimY)*k + i+1+l] - u_upd[(dimX*dimY)*k + i+l]; + else + gradx_upd[(dimX*dimY)*k + i+l] = 0.0f; + + // Derivatives in the y-direction + if(j != dimY-1) + //grady_upd[(dimX*dimY)*k + i+l] = u_upd[(dimX*dimY)*k + i+dimY+l] -u_upd[(dimX*dimY)*k + i+l]; + grady_upd[(dimX*dimY)*k + i+l] = u_upd[(dimX*dimY)*k + i+(j+1)*dimX] -u_upd[(dimX*dimY)*k + i+l]; + else + grady_upd[(dimX*dimY)*k + i+l] = 0.0f; + }}} + return 1; +} + +float proxF(float *gx, float *gy, float *vx, float *vy, float sigma, int p, int q, int r, long dimX, long dimY, long dimZ) +{ + // (S^p, \ell^1) norm decouples at each pixel +// Spl1(gx, gy, vx, vy, sigma, p, num_channels, dim); + float divsigma = 1.0f / sigma; + + // $\ell^{1,1,1}$-TV regularization +// int i,j,k; +// #pragma omp parallel for shared (gx,gy,vx,vy) private(i,j,k) +// for(k = 0; k < dimZ; k++) { +// for(i=0; i<dimX; i++) { +// for(j=0; j<dimY; j++) { +// gx[(dimX*dimY)*k + (i)*dimY + (j)] = SIGN(vx[(dimX*dimY)*k + (i)*dimY + (j)]) * MAX(fabs(vx[(dimX*dimY)*k + (i)*dimY + (j)]) - divsigma, 0.0f); +// gy[(dimX*dimY)*k + (i)*dimY + (j)] = SIGN(vy[(dimX*dimY)*k + (i)*dimY + (j)]) * MAX(fabs(vy[(dimX*dimY)*k + (i)*dimY + (j)]) - divsigma, 0.0f); +// }}} + + // Auxiliar vector + float *proj, sum, shrinkfactor ; + float M1,M2,M3,valuex,valuey,T,D,det,eig1,eig2,sig1,sig2,V1, V2, V3, V4, v0,v1,v2, mu1,mu2,sig1_upd,sig2_upd,t1,t2,t3; + long i,j,k, ii, num; +#pragma omp parallel for shared (gx,gy,vx,vy,p) private(i,ii,j,k,proj,num, sum, shrinkfactor, M1,M2,M3,valuex,valuey,T,D,det,eig1,eig2,sig1,sig2,V1, V2, V3, V4,v0,v1,v2,mu1,mu2,sig1_upd,sig2_upd,t1,t2,t3) + for(i=0; i<dimX; i++) { + for(j=0; j<dimY; j++) { + + proj = (float*) calloc (2,sizeof(float)); + // Compute matrix $M\in\R^{2\times 2}$ + M1 = 0.0f; + M2 = 0.0f; + M3 = 0.0f; + + for(k = 0; k < dimZ; k++) + { + valuex = vx[(dimX*dimY)*k + (j)*dimX + (i)]; + valuey = vy[(dimX*dimY)*k + (j)*dimX + (i)]; + + M1 += (valuex * valuex); + M2 += (valuex * valuey); + M3 += (valuey * valuey); + } + + // Compute eigenvalues of M + T = M1 + M3; + D = M1 * M3 - M2 * M2; + det = sqrt(MAX((T * T / 4.0f) - D, 0.0f)); + eig1 = MAX((T / 2.0f) + det, 0.0f); + eig2 = MAX((T / 2.0f) - det, 0.0f); + sig1 = sqrt(eig1); + sig2 = sqrt(eig2); + + // Compute normalized eigenvectors + V1 = V2 = V3 = V4 = 0.0f; + + if(M2 != 0.0f) + { + v0 = M2; + v1 = eig1 - M3; + v2 = eig2 - M3; + + mu1 = sqrtf(v0 * v0 + v1 * v1); + mu2 = sqrtf(v0 * v0 + v2 * v2); + + if(mu1 > fTiny) + { + V1 = v1 / mu1; + V3 = v0 / mu1; + } + + if(mu2 > fTiny) + { + V2 = v2 / mu2; + V4 = v0 / mu2; + } + + } else + { + if(M1 > M3) + { + V1 = V4 = 1.0f; + V2 = V3 = 0.0f; + + } else + { + V1 = V4 = 0.0f; + V2 = V3 = 1.0f; + } + } + + // Compute prox_p of the diagonal entries + sig1_upd = sig2_upd = 0.0f; + + if(p == 1) + { + sig1_upd = MAX(sig1 - divsigma, 0.0f); + sig2_upd = MAX(sig2 - divsigma, 0.0f); + + } else if(p == INFNORM) + { + proj[0] = sigma * fabs(sig1); + proj[1] = sigma * fabs(sig2); + + /*l1 projection part */ + sum = fLarge; + num = 0l; + shrinkfactor = 0.0f; + while(sum > 1.0f) + { + sum = 0.0f; + num = 0; + + for(ii = 0; ii < 2; ii++) + { + proj[ii] = MAX(proj[ii] - shrinkfactor, 0.0f); + + sum += fabs(proj[ii]); + if(proj[ii]!= 0.0f) + num++; + } + + if(num > 0) + shrinkfactor = (sum - 1.0f) / num; + else + break; + } + /*l1 proj ends*/ + + sig1_upd = sig1 - divsigma * proj[0]; + sig2_upd = sig2 - divsigma * proj[1]; + } + + // Compute the diagonal entries of $\widehat{\Sigma}\Sigma^{\dagger}_0$ + if(sig1 > fTiny) + sig1_upd /= sig1; + + if(sig2 > fTiny) + sig2_upd /= sig2; + + // Compute solution + t1 = sig1_upd * V1 * V1 + sig2_upd * V2 * V2; + t2 = sig1_upd * V1 * V3 + sig2_upd * V2 * V4; + t3 = sig1_upd * V3 * V3 + sig2_upd * V4 * V4; + + for(k = 0; k < dimZ; k++) + { + gx[(dimX*dimY)*k + j*dimX + i] = vx[(dimX*dimY)*k + j*dimX + i] * t1 + vy[(dimX*dimY)*k + j*dimX + i] * t2; + gy[(dimX*dimY)*k + j*dimX + i] = vx[(dimX*dimY)*k + j*dimX + i] * t2 + vy[(dimX*dimY)*k + j*dimX + i] * t3; + } + + // Delete allocated memory + free(proj); + }} + + return 1; +} + +float divergence(float *qx_upd, float *qy_upd, float *div_upd, long dimX, long dimY, long dimZ) +{ + long i, j, k, l; +#pragma omp parallel for shared(qx_upd,qy_upd,div_upd) private(i, j, k, l) + for(k = 0; k < dimZ; k++) { + for(j = 0; j < dimY; j++) { + l = j * dimX; + for(i = 0; i < dimX; i++) { + if(i != dimX-1) + { + // ux[k][i+l] = u[k][i+1+l] - u[k][i+l] + div_upd[(dimX*dimY)*k + i+1+l] -= qx_upd[(dimX*dimY)*k + i+l]; + div_upd[(dimX*dimY)*k + i+l] += qx_upd[(dimX*dimY)*k + i+l]; + } + + if(j != dimY-1) + { + // uy[k][i+l] = u[k][i+width+l] - u[k][i+l] + //div_upd[(dimX*dimY)*k + i+dimY+l] -= qy_upd[(dimX*dimY)*k + i+l]; + div_upd[(dimX*dimY)*k + i+(j+1)*dimX] -= qy_upd[(dimX*dimY)*k + i+l]; + div_upd[(dimX*dimY)*k + i+l] += qy_upd[(dimX*dimY)*k + i+l]; + } + } + } + } + return *div_upd; +} diff --git a/src/Core/regularisers_CPU/TNV_core.h b/src/Core/regularisers_CPU/TNV_core.h new file mode 100644 index 0000000..aa050a4 --- /dev/null +++ b/src/Core/regularisers_CPU/TNV_core.h @@ -0,0 +1,47 @@ +#include <math.h> +#include <stdlib.h> +#include <memory.h> +#include <stdio.h> +#include "omp.h" +#include "utils.h" +#include "CCPiDefines.h" + +#define fTiny 0.00000001f +#define fLarge 100000000.0f +#define INFNORM -1 + +#define MAX(i,j) ((i)<(j) ? (j):(i)) +#define MIN(i,j) ((i)<(j) ? (i):(j)) + +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float TNV_CPU_main(float *Input, float *u, float lambda, int maxIter, float tol, int dimX, int dimY, int dimZ); + +/*float PDHG(float *A, float *B, float tau, float sigma, float theta, float lambda, int p, int q, int r, float tol, int maxIter, int d_c, int d_w, int d_h);*/ +CCPI_EXPORT float proxG(float *u_upd, float *v, float *f, float taulambda, long dimX, long dimY, long dimZ); +CCPI_EXPORT float gradient(float *u_upd, float *gradx_upd, float *grady_upd, long dimX, long dimY, long dimZ); +CCPI_EXPORT float proxF(float *gx, float *gy, float *vx, float *vy, float sigma, int p, int q, int r, long dimX, long dimY, long dimZ); +CCPI_EXPORT float divergence(float *qx_upd, float *qy_upd, float *div_upd, long dimX, long dimY, long dimZ); +#ifdef __cplusplus +} +#endif
\ No newline at end of file diff --git a/src/Core/regularisers_CPU/utils.c b/src/Core/regularisers_CPU/utils.c new file mode 100644 index 0000000..7a4e80b --- /dev/null +++ b/src/Core/regularisers_CPU/utils.c @@ -0,0 +1,117 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazanteev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "utils.h" +#include <math.h> + +/* Copy Image (float) */ +float copyIm(float *A, float *U, long dimX, long dimY, long dimZ) +{ + long j; +#pragma omp parallel for shared(A, U) private(j) + for (j = 0; j<dimX*dimY*dimZ; j++) U[j] = A[j]; + return *U; +} + +/* Copy Image */ +unsigned char copyIm_unchar(unsigned char *A, unsigned char *U, int dimX, int dimY, int dimZ) +{ + int j; +#pragma omp parallel for shared(A, U) private(j) + for (j = 0; j<dimX*dimY*dimZ; j++) U[j] = A[j]; + return *U; +} + +/*Roll image symmetrically from top to bottom*/ +float copyIm_roll(float *A, float *U, int dimX, int dimY, int roll_value, int switcher) +{ + int i, j; +#pragma omp parallel for shared(U, A) private(i,j) + for (i=0; i<dimX; i++) { + for (j=0; j<dimY; j++) { + if (switcher == 0) { + if (j < (dimY - roll_value)) U[j*dimX + i] = A[(j+roll_value)*dimX + i]; + else U[j*dimX + i] = A[(j - (dimY - roll_value))*dimX + i]; + } + else { + if (j < roll_value) U[j*dimX + i] = A[(j+(dimY - roll_value))*dimX + i]; + else U[j*dimX + i] = A[(j - roll_value)*dimX + i]; + } + }} + return *U; +} + +/* function that calculates TV energy + * type - 1: 2*lambda*min||\nabla u|| + ||u -u0||^2 + * type - 2: 2*lambda*min||\nabla u|| + * */ +float TV_energy2D(float *U, float *U0, float *E_val, float lambda, int type, int dimX, int dimY) +{ + int i, j, i1, j1, index; + float NOMx_2, NOMy_2, E_Grad=0.0f, E_Data=0.0f; + + /* first calculate \grad U_xy*/ + for(j=0; j<dimY; j++) { + for(i=0; i<dimX; i++) { + index = j*dimX+i; + /* boundary conditions */ + i1 = i + 1; if (i == dimX-1) i1 = i; + j1 = j + 1; if (j == dimY-1) j1 = j; + + /* Forward differences */ + NOMx_2 = powf((float)(U[j1*dimX + i] - U[index]),2); /* x+ */ + NOMy_2 = powf((float)(U[j*dimX + i1] - U[index]),2); /* y+ */ + E_Grad += 2.0f*lambda*sqrtf((float)(NOMx_2) + (float)(NOMy_2)); /* gradient term energy */ + E_Data += powf((float)(U[index]-U0[index]),2); /* fidelity term energy */ + } + } + if (type == 1) E_val[0] = E_Grad + E_Data; + if (type == 2) E_val[0] = E_Grad; + return *E_val; +} + +float TV_energy3D(float *U, float *U0, float *E_val, float lambda, int type, int dimX, int dimY, int dimZ) +{ + long i, j, k, i1, j1, k1, index; + float NOMx_2, NOMy_2, NOMz_2, E_Grad=0.0f, E_Data=0.0f; + + /* first calculate \grad U_xy*/ + for(j=0; j<(long)(dimY); j++) { + for(i=0; i<(long)(dimX); i++) { + for(k=0; k<(long)(dimZ); k++) { + index = (dimX*dimY)*k + j*dimX+i; + /* boundary conditions */ + i1 = i + 1; if (i == (long)(dimX-1)) i1 = i; + j1 = j + 1; if (j == (long)(dimY-1)) j1 = j; + k1 = k + 1; if (k == (long)(dimZ-1)) k1 = k; + + /* Forward differences */ + NOMx_2 = powf((float)(U[(dimX*dimY)*k + j1*dimX+i] - U[index]),2); /* x+ */ + NOMy_2 = powf((float)(U[(dimX*dimY)*k + j*dimX+i1] - U[index]),2); /* y+ */ + NOMz_2 = powf((float)(U[(dimX*dimY)*k1 + j*dimX+i] - U[index]),2); /* z+ */ + + E_Grad += 2.0f*lambda*sqrtf((float)(NOMx_2) + (float)(NOMy_2) + (float)(NOMz_2)); /* gradient term energy */ + E_Data += (powf((float)(U[index]-U0[index]),2)); /* fidelity term energy */ + } + } + } + if (type == 1) E_val[0] = E_Grad + E_Data; + if (type == 2) E_val[0] = E_Grad; + return *E_val; +} diff --git a/src/Core/regularisers_CPU/utils.h b/src/Core/regularisers_CPU/utils.h new file mode 100644 index 0000000..cfaf6d7 --- /dev/null +++ b/src/Core/regularisers_CPU/utils.h @@ -0,0 +1,34 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include <stdlib.h> +#include <memory.h> +#include "CCPiDefines.h" +#include "omp.h" +#ifdef __cplusplus +extern "C" { +#endif +CCPI_EXPORT float copyIm(float *A, float *U, long dimX, long dimY, long dimZ); +CCPI_EXPORT unsigned char copyIm_unchar(unsigned char *A, unsigned char *U, int dimX, int dimY, int dimZ); +CCPI_EXPORT float copyIm_roll(float *A, float *U, int dimX, int dimY, int roll_value, int switcher); +CCPI_EXPORT float TV_energy2D(float *U, float *U0, float *E_val, float lambda, int type, int dimX, int dimY); +CCPI_EXPORT float TV_energy3D(float *U, float *U0, float *E_val, float lambda, int type, int dimX, int dimY, int dimZ); +#ifdef __cplusplus +} +#endif diff --git a/src/Core/regularisers_GPU/Diffus_4thO_GPU_core.cu b/src/Core/regularisers_GPU/Diffus_4thO_GPU_core.cu new file mode 100644 index 0000000..a4dbe70 --- /dev/null +++ b/src/Core/regularisers_GPU/Diffus_4thO_GPU_core.cu @@ -0,0 +1,268 @@ + /* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "Diffus_4thO_GPU_core.h" +#include "shared.h" + +/* CUDA implementation of fourth-order diffusion scheme [1] for piecewise-smooth recovery (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambda - regularization parameter + * 3. Edge-preserving parameter (sigma) + * 4. Number of iterations, for explicit scheme >= 150 is recommended + * 5. tau - time-marching step for explicit scheme + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Hajiaboli, M.R., 2011. An anisotropic fourth-order diffusion filter for image noise removal. International Journal of Computer Vision, 92(2), pp.177-191. + */ + +#define BLKXSIZE 8 +#define BLKYSIZE 8 +#define BLKZSIZE 8 + +#define BLKXSIZE2D 16 +#define BLKYSIZE2D 16 +#define EPS 1.0e-7 +#define idivup(a, b) ( ((a)%(b) != 0) ? (a)/(b)+1 : (a)/(b) ) +/********************************************************************/ +/***************************2D Functions*****************************/ +/********************************************************************/ +__global__ void Weighted_Laplc2D_kernel(float *W_Lapl, float *U0, float sigma, int dimX, int dimY) +{ + int i1,i2,j1,j2; + float gradX, gradX_sq, gradY, gradY_sq, gradXX, gradYY, gradXY, xy_2, denom, V_norm, V_orth, c, c_sq; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + + gradX = 0.5f*(U0[j*dimX+i2] - U0[j*dimX+i1]); + gradX_sq = powf(gradX,2); + + gradY = 0.5f*(U0[j2*dimX+i] - U0[j1*dimX+i]); + gradY_sq = powf(gradY,2); + + gradXX = U0[j*dimX+i2] + U0[j*dimX+i1] - 2*U0[index]; + gradYY = U0[j2*dimX+i] + U0[j1*dimX+i] - 2*U0[index]; + + gradXY = 0.25f*(U0[j2*dimX+i2] + U0[j1*dimX+i1] - U0[j1*dimX+i2] - U0[j2*dimX+i1]); + xy_2 = 2.0f*gradX*gradY*gradXY; + + denom = gradX_sq + gradY_sq; + + if (denom <= EPS) { + V_norm = (gradXX*gradX_sq + xy_2 + gradYY*gradY_sq)/EPS; + V_orth = (gradXX*gradY_sq - xy_2 + gradYY*gradX_sq)/EPS; + } + else { + V_norm = (gradXX*gradX_sq + xy_2 + gradYY*gradY_sq)/denom; + V_orth = (gradXX*gradY_sq - xy_2 + gradYY*gradX_sq)/denom; + } + + c = 1.0f/(1.0f + denom/sigma); + c_sq = c*c; + + W_Lapl[index] = c_sq*V_norm + c*V_orth; + } + return; +} + +__global__ void Diffusion_update_step2D_kernel(float *Output, float *Input, float *W_Lapl, float lambdaPar, float sigmaPar2, float tau, int dimX, int dimY) +{ + int i1,i2,j1,j2; + float gradXXc, gradYYc; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + + gradXXc = W_Lapl[j*dimX+i2] + W_Lapl[j*dimX+i1] - 2*W_Lapl[index]; + gradYYc = W_Lapl[j2*dimX+i] + W_Lapl[j1*dimX+i] - 2*W_Lapl[index]; + + Output[index] += tau*(-lambdaPar*(gradXXc + gradYYc) - (Output[index] - Input[index])); + } + return; +} +/********************************************************************/ +/***************************3D Functions*****************************/ +/********************************************************************/ +__global__ void Weighted_Laplc3D_kernel(float *W_Lapl, float *U0, float sigma, int dimX, int dimY, int dimZ) +{ + int i1,i2,j1,j2,k1,k2; + float gradX, gradX_sq, gradY, gradY_sq, gradXX, gradYY, gradXY, xy_2, denom, V_norm, V_orth, c, c_sq, gradZ, gradZ_sq, gradZZ, gradXZ, gradYZ, xyz_1, xyz_2; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + k1 = k+1; if (k1 == dimZ) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + + int index = (dimX*dimY)*k + j*dimX+i; + + gradX = 0.5f*(U0[(dimX*dimY)*k + j*dimX+i2] - U0[(dimX*dimY)*k + j*dimX+i1]); + gradX_sq = pow(gradX,2); + + gradY = 0.5f*(U0[(dimX*dimY)*k + j2*dimX+i] - U0[(dimX*dimY)*k + j1*dimX+i]); + gradY_sq = pow(gradY,2); + + gradZ = 0.5f*(U0[(dimX*dimY)*k2 + j*dimX+i] - U0[(dimX*dimY)*k1 + j*dimX+i]); + gradZ_sq = pow(gradZ,2); + + gradXX = U0[(dimX*dimY)*k + j*dimX+i2] + U0[(dimX*dimY)*k + j*dimX+i1] - 2*U0[index]; + gradYY = U0[(dimX*dimY)*k + j2*dimX+i] + U0[(dimX*dimY)*k + j1*dimX+i] - 2*U0[index]; + gradZZ = U0[(dimX*dimY)*k2 + j*dimX+i] + U0[(dimX*dimY)*k1 + j*dimX+i] - 2*U0[index]; + + gradXY = 0.25f*(U0[(dimX*dimY)*k + j2*dimX+i2] + U0[(dimX*dimY)*k + j1*dimX+i1] - U0[(dimX*dimY)*k + j1*dimX+i2] - U0[(dimX*dimY)*k + j2*dimX+i1]); + gradXZ = 0.25f*(U0[(dimX*dimY)*k2 + j*dimX+i2] - U0[(dimX*dimY)*k2+j*dimX+i1] - U0[(dimX*dimY)*k1+j*dimX+i2] + U0[(dimX*dimY)*k1+j*dimX+i1]); + gradYZ = 0.25f*(U0[(dimX*dimY)*k2 +j2*dimX+i] - U0[(dimX*dimY)*k2+j1*dimX+i] - U0[(dimX*dimY)*k1+j2*dimX+i] + U0[(dimX*dimY)*k1+j1*dimX+i]); + + xy_2 = 2.0f*gradX*gradY*gradXY; + xyz_1 = 2.0f*gradX*gradZ*gradXZ; + xyz_2 = 2.0f*gradY*gradZ*gradYZ; + + denom = gradX_sq + gradY_sq + gradZ_sq; + + if (denom <= EPS) { + V_norm = (gradXX*gradX_sq + gradYY*gradY_sq + gradZZ*gradZ_sq + xy_2 + xyz_1 + xyz_2)/EPS; + V_orth = ((gradY_sq + gradZ_sq)*gradXX + (gradX_sq + gradZ_sq)*gradYY + (gradX_sq + gradY_sq)*gradZZ - xy_2 - xyz_1 - xyz_2)/EPS; + } + else { + V_norm = (gradXX*gradX_sq + gradYY*gradY_sq + gradZZ*gradZ_sq + xy_2 + xyz_1 + xyz_2)/denom; + V_orth = ((gradY_sq + gradZ_sq)*gradXX + (gradX_sq + gradZ_sq)*gradYY + (gradX_sq + gradY_sq)*gradZZ - xy_2 - xyz_1 - xyz_2)/denom; + } + + c = 1.0f/(1.0f + denom/sigma); + c_sq = c*c; + + W_Lapl[index] = c_sq*V_norm + c*V_orth; + } + return; +} +__global__ void Diffusion_update_step3D_kernel(float *Output, float *Input, float *W_Lapl, float lambdaPar, float sigmaPar2, float tau, int dimX, int dimY, int dimZ) +{ + int i1,i2,j1,j2,k1,k2; + float gradXXc, gradYYc, gradZZc; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i+1; if (i1 == dimX) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == dimY) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + k1 = k+1; if (k1 == dimZ) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + + int index = (dimX*dimY)*k + j*dimX+i; + + gradXXc = W_Lapl[(dimX*dimY)*k + j*dimX+i2] + W_Lapl[(dimX*dimY)*k + j*dimX+i1] - 2*W_Lapl[index]; + gradYYc = W_Lapl[(dimX*dimY)*k + j2*dimX+i] + W_Lapl[(dimX*dimY)*k + j1*dimX+i] - 2*W_Lapl[index]; + gradZZc = W_Lapl[(dimX*dimY)*k2 + j*dimX+i] + W_Lapl[(dimX*dimY)*k1 + j*dimX+i] - 2*W_Lapl[index]; + + Output[index] += tau*(-lambdaPar*(gradXXc + gradYYc + gradZZc) - (Output[index] - Input[index])); + } + return; +} +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ +/********************* MAIN HOST FUNCTION ******************/ +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ +extern "C" int Diffus4th_GPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int N, int M, int Z) +{ + int dimTotal, dev = 0; + CHECK(cudaSetDevice(dev)); + float *d_input, *d_output, *d_W_Lapl; + float sigmaPar2; + sigmaPar2 = sigmaPar*sigmaPar; + dimTotal = N*M*Z; + + CHECK(cudaMalloc((void**)&d_input,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&d_output,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&d_W_Lapl,dimTotal*sizeof(float))); + + CHECK(cudaMemcpy(d_input,Input,dimTotal*sizeof(float),cudaMemcpyHostToDevice)); + CHECK(cudaMemcpy(d_output,Input,dimTotal*sizeof(float),cudaMemcpyHostToDevice)); + + if (Z == 1) { + /*2D case */ + dim3 dimBlock(BLKXSIZE2D,BLKYSIZE2D); + dim3 dimGrid(idivup(N,BLKXSIZE2D), idivup(M,BLKYSIZE2D)); + + for(int n=0; n < iterationsNumb; n++) { + /* Calculating weighted Laplacian */ + Weighted_Laplc2D_kernel<<<dimGrid,dimBlock>>>(d_W_Lapl, d_output, sigmaPar2, N, M); + CHECK(cudaDeviceSynchronize()); + /* Perform iteration step */ + Diffusion_update_step2D_kernel<<<dimGrid,dimBlock>>>(d_output, d_input, d_W_Lapl, lambdaPar, sigmaPar2, tau, N, M); + CHECK(cudaDeviceSynchronize()); + } + } + else { + /*3D case*/ + dim3 dimBlock(BLKXSIZE,BLKYSIZE,BLKZSIZE); + dim3 dimGrid(idivup(N,BLKXSIZE), idivup(M,BLKYSIZE),idivup(Z,BLKZSIZE)); + for(int n=0; n < iterationsNumb; n++) { + /* Calculating weighted Laplacian */ + Weighted_Laplc3D_kernel<<<dimGrid,dimBlock>>>(d_W_Lapl, d_output, sigmaPar2, N, M, Z); + CHECK(cudaDeviceSynchronize()); + /* Perform iteration step */ + Diffusion_update_step3D_kernel<<<dimGrid,dimBlock>>>(d_output, d_input, d_W_Lapl, lambdaPar, sigmaPar2, tau, N, M, Z); + CHECK(cudaDeviceSynchronize()); + } + } + CHECK(cudaMemcpy(Output,d_output,dimTotal*sizeof(float),cudaMemcpyDeviceToHost)); + CHECK(cudaFree(d_input)); + CHECK(cudaFree(d_output)); + CHECK(cudaFree(d_W_Lapl)); + return 0; +} diff --git a/src/Core/regularisers_GPU/Diffus_4thO_GPU_core.h b/src/Core/regularisers_GPU/Diffus_4thO_GPU_core.h new file mode 100644 index 0000000..77d5d79 --- /dev/null +++ b/src/Core/regularisers_GPU/Diffus_4thO_GPU_core.h @@ -0,0 +1,8 @@ +#ifndef __Diff_4thO_GPU_H__ +#define __Diff_4thO_GPU_H__ +#include "CCPiDefines.h" +#include <stdio.h> + +extern "C" CCPI_EXPORT int Diffus4th_GPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int N, int M, int Z); + +#endif diff --git a/src/Core/regularisers_GPU/LLT_ROF_GPU_core.cu b/src/Core/regularisers_GPU/LLT_ROF_GPU_core.cu new file mode 100644 index 0000000..87871be --- /dev/null +++ b/src/Core/regularisers_GPU/LLT_ROF_GPU_core.cu @@ -0,0 +1,473 @@ + /* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "LLT_ROF_GPU_core.h" +#include "shared.h" + +/* CUDA implementation of Lysaker, Lundervold and Tai (LLT) model [1] combined with Rudin-Osher-Fatemi [2] TV regularisation penalty. + * +* This penalty can deliver visually pleasant piecewise-smooth recovery if regularisation parameters are selected well. +* The rule of thumb for selection is to start with lambdaLLT = 0 (just the ROF-TV model) and then proceed to increase +* lambdaLLT starting with smaller values. +* +* Input Parameters: +* 1. U0 - original noise image/volume +* 2. lambdaROF - ROF-related regularisation parameter +* 3. lambdaLLT - LLT-related regularisation parameter +* 4. tau - time-marching step +* 5. iter - iterations number (for both models) +* +* Output: +* Filtered/regularised image +* +* References: +* [1] Lysaker, M., Lundervold, A. and Tai, X.C., 2003. Noise removal using fourth-order partial differential equation with applications to medical magnetic resonance images in space and time. IEEE Transactions on image processing, 12(12), pp.1579-1590. +* [2] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" +*/ + +#define BLKXSIZE 8 +#define BLKYSIZE 8 +#define BLKZSIZE 8 + +#define BLKXSIZE2D 16 +#define BLKYSIZE2D 16 + + +#define EPS_LLT 0.01 +#define EPS_ROF 1.0e-12 + +#define idivup(a, b) ( ((a)%(b) != 0) ? (a)/(b)+1 : (a)/(b) ) + +#define MAX(x, y) (((x) > (y)) ? (x) : (y)) +#define MIN(x, y) (((x) < (y)) ? (x) : (y)) + +__host__ __device__ int signLLT (float x) +{ + return (x > 0) - (x < 0); +} + +/*************************************************************************/ +/**********************LLT-related functions *****************************/ +/*************************************************************************/ +__global__ void der2D_LLT_kernel(float *U, float *D1, float *D2, int dimX, int dimY) + { + int i_p, i_m, j_m, j_p; + float dxx, dyy, denom_xx, denom_yy; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + + /* symmetric boundary conditions (Neuman) */ + i_p = i + 1; if (i_p == dimX) i_p = i - 1; + i_m = i - 1; if (i_m < 0) i_m = i + 1; + j_p = j + 1; if (j_p == dimY) j_p = j - 1; + j_m = j - 1; if (j_m < 0) j_m = j + 1; + + dxx = U[j*dimX+i_p] - 2.0f*U[index] + U[j*dimX+i_m]; + dyy = U[j_p*dimX+i] - 2.0f*U[index] + U[j_m*dimX+i]; + + denom_xx = abs(dxx) + EPS_LLT; + denom_yy = abs(dyy) + EPS_LLT; + + D1[index] = dxx / denom_xx; + D2[index] = dyy / denom_yy; + } + } + +__global__ void der3D_LLT_kernel(float* U, float *D1, float *D2, float *D3, int dimX, int dimY, int dimZ) + { + int i_p, i_m, j_m, j_p, k_p, k_m; + float dxx, dyy, dzz, denom_xx, denom_yy, denom_zz; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + /* symmetric boundary conditions (Neuman) */ + i_p = i + 1; if (i_p == dimX) i_p = i - 1; + i_m = i - 1; if (i_m < 0) i_m = i + 1; + j_p = j + 1; if (j_p == dimY) j_p = j - 1; + j_m = j - 1; if (j_m < 0) j_m = j + 1; + k_p = k + 1; if (k_p == dimZ) k_p = k - 1; + k_m = k - 1; if (k_m < 0) k_m = k + 1; + + int index = (dimX*dimY)*k + j*dimX+i; + + dxx = U[(dimX*dimY)*k + j*dimX+i_p] - 2.0f*U[index] + U[(dimX*dimY)*k + j*dimX+i_m]; + dyy = U[(dimX*dimY)*k + j_p*dimX+i] - 2.0f*U[index] + U[(dimX*dimY)*k + j_m*dimX+i]; + dzz = U[(dimX*dimY)*k_p + j*dimX+i] - 2.0f*U[index] + U[(dimX*dimY)*k_m + j*dimX+i]; + + denom_xx = abs(dxx) + EPS_LLT; + denom_yy = abs(dyy) + EPS_LLT; + denom_zz = abs(dzz) + EPS_LLT; + + D1[index] = dxx / denom_xx; + D2[index] = dyy / denom_yy; + D3[index] = dzz / denom_zz; + } + } + +/*************************************************************************/ +/**********************ROF-related functions *****************************/ +/*************************************************************************/ + +/* first-order differences 1 */ +__global__ void D1_func2D_ROF_kernel(float* Input, float* D1, int N, int M) + { + int i1, j1, i2; + float NOMx_1,NOMy_1,NOMy_0,denom1,denom2,T1; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + N*j; + + if ((i >= 0) && (i < N) && (j >= 0) && (j < M)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i + 1; if (i1 >= N) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= M) j1 = j-1; + + /* Forward-backward differences */ + NOMx_1 = Input[j1*N + i] - Input[index]; /* x+ */ + NOMy_1 = Input[j*N + i1] - Input[index]; /* y+ */ + NOMy_0 = Input[index] - Input[j*N + i2]; /* y- */ + + denom1 = NOMx_1*NOMx_1; + denom2 = 0.5f*(signLLT((float)NOMy_1) + signLLT((float)NOMy_0))*(MIN(abs((float)NOMy_1),abs((float)NOMy_0))); + denom2 = denom2*denom2; + T1 = sqrt(denom1 + denom2 + EPS_ROF); + D1[index] = NOMx_1/T1; + } + } + +/* differences 2 */ +__global__ void D2_func2D_ROF_kernel(float* Input, float* D2, int N, int M) + { + int i1, j1, j2; + float NOMx_1,NOMy_1,NOMx_0,denom1,denom2,T2; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + N*j; + + if ((i >= 0) && (i < (N)) && (j >= 0) && (j < (M))) { + + /* boundary conditions (Neumann reflections) */ + i1 = i + 1; if (i1 >= N) i1 = i-1; + j1 = j + 1; if (j1 >= M) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + + /* Forward-backward differences */ + NOMx_1 = Input[j1*N + i] - Input[index]; /* x+ */ + NOMy_1 = Input[j*N + i1] - Input[index]; /* y+ */ + NOMx_0 = Input[index] - Input[j2*N + i]; /* x- */ + + denom1 = NOMy_1*NOMy_1; + denom2 = 0.5f*(signLLT((float)NOMx_1) + signLLT((float)NOMx_0))*(MIN(abs((float)NOMx_1),abs((float)NOMx_0))); + denom2 = denom2*denom2; + T2 = sqrt(denom1 + denom2 + EPS_ROF); + D2[index] = NOMy_1/T2; + } + } + + + /* differences 1 */ +__global__ void D1_func3D_ROF_kernel(float* Input, float* D1, int dimX, int dimY, int dimZ) + { + float NOMx_1, NOMy_1, NOMy_0, NOMz_1, NOMz_0, denom1, denom2,denom3, T1; + int i1,i2,k1,j1,j2,k2; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (dimX*dimY)*k + j*dimX+i; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /* Forward-backward differences */ + NOMx_1 = Input[(dimX*dimY)*k + j1*dimX + i] - Input[index]; /* x+ */ + NOMy_1 = Input[(dimX*dimY)*k + j*dimX + i1] - Input[index]; /* y+ */ + NOMy_0 = Input[index] - Input[(dimX*dimY)*k + j*dimX + i2]; /* y- */ + + NOMz_1 = Input[(dimX*dimY)*k1 + j*dimX + i] - Input[index]; /* z+ */ + NOMz_0 = Input[index] - Input[(dimX*dimY)*k2 + j*dimX + i]; /* z- */ + + + denom1 = NOMx_1*NOMx_1; + denom2 = 0.5*(signLLT(NOMy_1) + signLLT(NOMy_0))*(MIN(abs(NOMy_1),abs(NOMy_0))); + denom2 = denom2*denom2; + denom3 = 0.5*(signLLT(NOMz_1) + signLLT(NOMz_0))*(MIN(abs(NOMz_1),abs(NOMz_0))); + denom3 = denom3*denom3; + T1 = sqrt(denom1 + denom2 + denom3 + EPS_ROF); + D1[index] = NOMx_1/T1; + } + } + + /* differences 2 */ + __global__ void D2_func3D_ROF_kernel(float* Input, float* D2, int dimX, int dimY, int dimZ) + { + float NOMx_1, NOMy_1, NOMx_0, NOMz_1, NOMz_0, denom1, denom2, denom3, T2; + int i1,i2,k1,j1,j2,k2; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (dimX*dimY)*k + j*dimX+i; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + + /* Forward-backward differences */ + NOMx_1 = Input[(dimX*dimY)*k + (j1)*dimX + i] - Input[index]; /* x+ */ + NOMy_1 = Input[(dimX*dimY)*k + (j)*dimX + i1] - Input[index]; /* y+ */ + NOMx_0 = Input[index] - Input[(dimX*dimY)*k + (j2)*dimX + i]; /* x- */ + NOMz_1 = Input[(dimX*dimY)*k1 + j*dimX + i] - Input[index]; /* z+ */ + NOMz_0 = Input[index] - Input[(dimX*dimY)*k2 + (j)*dimX + i]; /* z- */ + + + denom1 = NOMy_1*NOMy_1; + denom2 = 0.5*(signLLT(NOMx_1) + signLLT(NOMx_0))*(MIN(abs(NOMx_1),abs(NOMx_0))); + denom2 = denom2*denom2; + denom3 = 0.5*(signLLT(NOMz_1) + signLLT(NOMz_0))*(MIN(abs(NOMz_1),abs(NOMz_0))); + denom3 = denom3*denom3; + T2 = sqrt(denom1 + denom2 + denom3 + EPS_ROF); + D2[index] = NOMy_1/T2; + } + } + + /* differences 3 */ + __global__ void D3_func3D_ROF_kernel(float* Input, float* D3, int dimX, int dimY, int dimZ) + { + float NOMx_1, NOMy_1, NOMx_0, NOMy_0, NOMz_1, denom1, denom2, denom3, T3; + int i1,i2,k1,j1,j2,k2; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (dimX*dimY)*k + j*dimX+i; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /* Forward-backward differences */ + NOMx_1 = Input[(dimX*dimY)*k + (j1)*dimX + i] - Input[index]; /* x+ */ + NOMy_1 = Input[(dimX*dimY)*k + (j)*dimX + i1] - Input[index]; /* y+ */ + NOMy_0 = Input[index] - Input[(dimX*dimY)*k + (j)*dimX + i2]; /* y- */ + NOMx_0 = Input[index] - Input[(dimX*dimY)*k + (j2)*dimX + i]; /* x- */ + NOMz_1 = Input[(dimX*dimY)*k1 + j*dimX + i] - Input[index]; /* z+ */ + + denom1 = NOMz_1*NOMz_1; + denom2 = 0.5*(signLLT(NOMx_1) + signLLT(NOMx_0))*(MIN(abs(NOMx_1),abs(NOMx_0))); + denom2 = denom2*denom2; + denom3 = 0.5*(signLLT(NOMy_1) + signLLT(NOMy_0))*(MIN(abs(NOMy_1),abs(NOMy_0))); + denom3 = denom3*denom3; + T3 = sqrt(denom1 + denom2 + denom3 + EPS_ROF); + D3[index] = NOMz_1/T3; + } + } +/*************************************************************************/ +/**********************ROF-LLT-related functions *************************/ +/*************************************************************************/ + +__global__ void Update2D_LLT_ROF_kernel(float *U0, float *U, float *D1_LLT, float *D2_LLT, float *D1_ROF, float *D2_ROF, float lambdaROF, float lambdaLLT, float tau, int dimX, int dimY) +{ + + int i_p, i_m, j_m, j_p; + float div, laplc, dxx, dyy, dv1, dv2; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + + /* symmetric boundary conditions (Neuman) */ + i_p = i + 1; if (i_p == dimX) i_p = i - 1; + i_m = i - 1; if (i_m < 0) i_m = i + 1; + j_p = j + 1; if (j_p == dimY) j_p = j - 1; + j_m = j - 1; if (j_m < 0) j_m = j + 1; + + index = j*dimX+i; + + /*LLT-related part*/ + dxx = D1_LLT[j*dimX+i_p] - 2.0f*D1_LLT[index] + D1_LLT[j*dimX+i_m]; + dyy = D2_LLT[j_p*dimX+i] - 2.0f*D2_LLT[index] + D2_LLT[j_m*dimX+i]; + laplc = dxx + dyy; /*build Laplacian*/ + /*ROF-related part*/ + dv1 = D1_ROF[index] - D1_ROF[j_m*dimX + i]; + dv2 = D2_ROF[index] - D2_ROF[j*dimX + i_m]; + div = dv1 + dv2; /*build Divirgent*/ + + /*combine all into one cost function to minimise */ + U[index] += tau*(2.0f*lambdaROF*(div) - lambdaLLT*(laplc) - (U[index] - U0[index])); + } +} + +__global__ void Update3D_LLT_ROF_kernel(float *U0, float *U, float *D1_LLT, float *D2_LLT, float *D3_LLT, float *D1_ROF, float *D2_ROF, float *D3_ROF, float lambdaROF, float lambdaLLT, float tau, int dimX, int dimY, int dimZ) +{ + int i_p, i_m, j_m, j_p, k_p, k_m; + float div, laplc, dxx, dyy, dzz, dv1, dv2, dv3; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + /* symmetric boundary conditions (Neuman) */ + i_p = i + 1; if (i_p == dimX) i_p = i - 1; + i_m = i - 1; if (i_m < 0) i_m = i + 1; + j_p = j + 1; if (j_p == dimY) j_p = j - 1; + j_m = j - 1; if (j_m < 0) j_m = j + 1; + k_p = k + 1; if (k_p == dimZ) k_p = k - 1; + k_m = k - 1; if (k_m < 0) k_m = k + 1; + + int index = (dimX*dimY)*k + j*dimX+i; + + /*LLT-related part*/ + dxx = D1_LLT[(dimX*dimY)*k + j*dimX+i_p] - 2.0f*D1_LLT[index] + D1_LLT[(dimX*dimY)*k + j*dimX+i_m]; + dyy = D2_LLT[(dimX*dimY)*k + j_p*dimX+i] - 2.0f*D2_LLT[index] + D2_LLT[(dimX*dimY)*k + j_m*dimX+i]; + dzz = D3_LLT[(dimX*dimY)*k_p + j*dimX+i] - 2.0f*D3_LLT[index] + D3_LLT[(dimX*dimY)*k_m + j*dimX+i]; + laplc = dxx + dyy + dzz; /*build Laplacian*/ + + /*ROF-related part*/ + dv1 = D1_ROF[index] - D1_ROF[(dimX*dimY)*k + j_m*dimX+i]; + dv2 = D2_ROF[index] - D2_ROF[(dimX*dimY)*k + j*dimX+i_m]; + dv3 = D3_ROF[index] - D3_ROF[(dimX*dimY)*k_m + j*dimX+i]; + div = dv1 + dv2 + dv3; /*build Divirgent*/ + + /*combine all into one cost function to minimise */ + U[index] += tau*(2.0f*lambdaROF*(div) - lambdaLLT*(laplc) - (U[index] - U0[index])); + } +} + +/*******************************************************************/ +/************************ HOST FUNCTION ****************************/ +/*******************************************************************/ + +extern "C" int LLT_ROF_GPU_main(float *Input, float *Output, float lambdaROF, float lambdaLLT, int iterationsNumb, float tau, int N, int M, int Z) +{ + // set up device + int dev = 0; + int DimTotal; + DimTotal = N*M*Z; + CHECK(cudaSetDevice(dev)); + float *d_input, *d_update; + float *D1_LLT=NULL, *D2_LLT=NULL, *D3_LLT=NULL, *D1_ROF=NULL, *D2_ROF=NULL, *D3_ROF=NULL; + + if (Z == 0) {Z = 1;} + + CHECK(cudaMalloc((void**)&d_input,DimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&d_update,DimTotal*sizeof(float))); + + CHECK(cudaMalloc((void**)&D1_LLT,DimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&D2_LLT,DimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&D3_LLT,DimTotal*sizeof(float))); + + CHECK(cudaMalloc((void**)&D1_ROF,DimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&D2_ROF,DimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&D3_ROF,DimTotal*sizeof(float))); + + CHECK(cudaMemcpy(d_input,Input,DimTotal*sizeof(float),cudaMemcpyHostToDevice)); + CHECK(cudaMemcpy(d_update,Input,DimTotal*sizeof(float),cudaMemcpyHostToDevice)); + + if (Z == 1) { + // TV - 2D case + dim3 dimBlock(BLKXSIZE2D,BLKYSIZE2D); + dim3 dimGrid(idivup(N,BLKXSIZE2D), idivup(M,BLKYSIZE2D)); + + for(int n=0; n < iterationsNumb; n++) { + /****************ROF******************/ + /* calculate first-order differences */ + D1_func2D_ROF_kernel<<<dimGrid,dimBlock>>>(d_update, D1_ROF, N, M); + CHECK(cudaDeviceSynchronize()); + D2_func2D_ROF_kernel<<<dimGrid,dimBlock>>>(d_update, D2_ROF, N, M); + CHECK(cudaDeviceSynchronize()); + /****************LLT******************/ + /* estimate second-order derrivatives */ + der2D_LLT_kernel<<<dimGrid,dimBlock>>>(d_update, D1_LLT, D2_LLT, N, M); + /* Joint update for ROF and LLT models */ + Update2D_LLT_ROF_kernel<<<dimGrid,dimBlock>>>(d_input, d_update, D1_LLT, D2_LLT, D1_ROF, D2_ROF, lambdaROF, lambdaLLT, tau, N, M); + CHECK(cudaDeviceSynchronize()); + } + } + else { + // 3D case + dim3 dimBlock(BLKXSIZE,BLKYSIZE,BLKZSIZE); + dim3 dimGrid(idivup(N,BLKXSIZE), idivup(M,BLKYSIZE),idivup(Z,BLKXSIZE)); + + for(int n=0; n < iterationsNumb; n++) { + /****************ROF******************/ + /* calculate first-order differences */ + D1_func3D_ROF_kernel<<<dimGrid,dimBlock>>>(d_update, D1_ROF, N, M, Z); + CHECK(cudaDeviceSynchronize()); + D2_func3D_ROF_kernel<<<dimGrid,dimBlock>>>(d_update, D2_ROF, N, M, Z); + CHECK(cudaDeviceSynchronize()); + D3_func3D_ROF_kernel<<<dimGrid,dimBlock>>>(d_update, D3_ROF, N, M, Z); + CHECK(cudaDeviceSynchronize()); + /****************LLT******************/ + /* estimate second-order derrivatives */ + der3D_LLT_kernel<<<dimGrid,dimBlock>>>(d_update, D1_LLT, D2_LLT, D3_LLT, N, M, Z); + /* Joint update for ROF and LLT models */ + Update3D_LLT_ROF_kernel<<<dimGrid,dimBlock>>>(d_input, d_update, D1_LLT, D2_LLT, D3_LLT, D1_ROF, D2_ROF, D3_ROF, lambdaROF, lambdaLLT, tau, N, M, Z); + CHECK(cudaDeviceSynchronize()); + } + } + CHECK(cudaMemcpy(Output,d_update,DimTotal*sizeof(float),cudaMemcpyDeviceToHost)); + CHECK(cudaFree(d_input)); + CHECK(cudaFree(d_update)); + CHECK(cudaFree(D1_LLT)); + CHECK(cudaFree(D2_LLT)); + CHECK(cudaFree(D3_LLT)); + CHECK(cudaFree(D1_ROF)); + CHECK(cudaFree(D2_ROF)); + CHECK(cudaFree(D3_ROF)); + return 0; +} diff --git a/src/Core/regularisers_GPU/LLT_ROF_GPU_core.h b/src/Core/regularisers_GPU/LLT_ROF_GPU_core.h new file mode 100644 index 0000000..a6bfcc7 --- /dev/null +++ b/src/Core/regularisers_GPU/LLT_ROF_GPU_core.h @@ -0,0 +1,8 @@ +#ifndef __ROFLLTGPU_H__ +#define __ROFLLTGPU_H__ +#include "CCPiDefines.h" +#include <stdio.h> + +extern "C" CCPI_EXPORT int LLT_ROF_GPU_main(float *Input, float *Output, float lambdaROF, float lambdaLLT, int iterationsNumb, float tau, int N, int M, int Z); + +#endif diff --git a/src/Core/regularisers_GPU/NonlDiff_GPU_core.cu b/src/Core/regularisers_GPU/NonlDiff_GPU_core.cu new file mode 100644 index 0000000..ff7ce4d --- /dev/null +++ b/src/Core/regularisers_GPU/NonlDiff_GPU_core.cu @@ -0,0 +1,345 @@ + /* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "NonlDiff_GPU_core.h" +#include "shared.h" + +/* CUDA implementation of linear and nonlinear diffusion with the regularisation model [1,2] (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambda - regularization parameter + * 3. Edge-preserving parameter (sigma), when sigma equals to zero nonlinear diffusion -> linear diffusion + * 4. Number of iterations, for explicit scheme >= 150 is recommended + * 5. tau - time-marching step for explicit scheme + * 6. Penalty type: 1 - Huber, 2 - Perona-Malik, 3 - Tukey Biweight + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Perona, P. and Malik, J., 1990. Scale-space and edge detection using anisotropic diffusion. IEEE Transactions on pattern analysis and machine intelligence, 12(7), pp.629-639. + * [2] Black, M.J., Sapiro, G., Marimont, D.H. and Heeger, D., 1998. Robust anisotropic diffusion. IEEE Transactions on image processing, 7(3), pp.421-432. + */ + + +#define BLKXSIZE 8 +#define BLKYSIZE 8 +#define BLKZSIZE 8 + +#define BLKXSIZE2D 16 +#define BLKYSIZE2D 16 +#define EPS 1.0e-5 + +#define idivup(a, b) ( ((a)%(b) != 0) ? (a)/(b)+1 : (a)/(b) ) + +#define MAX(x, y) (((x) > (y)) ? (x) : (y)) +#define MIN(x, y) (((x) < (y)) ? (x) : (y)) + +__host__ __device__ int signNDF (float x) +{ + return (x > 0) - (x < 0); +} + +/********************************************************************/ +/***************************2D Functions*****************************/ +/********************************************************************/ +__global__ void LinearDiff2D_kernel(float *Input, float *Output, float lambdaPar, float tau, int N, int M) + { + int i1,i2,j1,j2; + float e,w,n,s,e1,w1,n1,s1; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + N*j; + + if ((i >= 0) && (i < N) && (j >= 0) && (j < M)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i+1; if (i1 == N) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == M) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + + e = Output[j*N+i1]; + w = Output[j*N+i2]; + n = Output[j1*N+i]; + s = Output[j2*N+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1) - (Output[index] - Input[index])); + } + } + + __global__ void NonLinearDiff2D_kernel(float *Input, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, int N, int M) + { + int i1,i2,j1,j2; + float e,w,n,s,e1,w1,n1,s1; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + N*j; + + if ((i >= 0) && (i < N) && (j >= 0) && (j < M)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i+1; if (i1 == N) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == M) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + + e = Output[j*N+i1]; + w = Output[j*N+i2]; + n = Output[j1*N+i]; + s = Output[j2*N+i]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + + if (penaltytype == 1){ + /* Huber penalty */ + if (abs(e1) > sigmaPar) e1 = signNDF(e1); + else e1 = e1/sigmaPar; + + if (abs(w1) > sigmaPar) w1 = signNDF(w1); + else w1 = w1/sigmaPar; + + if (abs(n1) > sigmaPar) n1 = signNDF(n1); + else n1 = n1/sigmaPar; + + if (abs(s1) > sigmaPar) s1 = signNDF(s1); + else s1 = s1/sigmaPar; + } + else if (penaltytype == 2) { + /* Perona-Malik */ + e1 = (e1)/(1.0f + pow((e1/sigmaPar),2)); + w1 = (w1)/(1.0f + pow((w1/sigmaPar),2)); + n1 = (n1)/(1.0f + pow((n1/sigmaPar),2)); + s1 = (s1)/(1.0f + pow((s1/sigmaPar),2)); + } + else if (penaltytype == 3) { + /* Tukey Biweight */ + if (abs(e1) <= sigmaPar) e1 = e1*pow((1.0f - pow((e1/sigmaPar),2)), 2); + else e1 = 0.0f; + if (abs(w1) <= sigmaPar) w1 = w1*pow((1.0f - pow((w1/sigmaPar),2)), 2); + else w1 = 0.0f; + if (abs(n1) <= sigmaPar) n1 = n1*pow((1.0f - pow((n1/sigmaPar),2)), 2); + else n1 = 0.0f; + if (abs(s1) <= sigmaPar) s1 = s1*pow((1.0f - pow((s1/sigmaPar),2)), 2); + else s1 = 0.0f; + } + else printf("%s \n", "No penalty function selected! Use 1,2 or 3."); + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1) - (Output[index] - Input[index])); + } + } +/********************************************************************/ +/***************************3D Functions*****************************/ +/********************************************************************/ + +__global__ void LinearDiff3D_kernel(float *Input, float *Output, float lambdaPar, float tau, int N, int M, int Z) + { + int i1,i2,j1,j2,k1,k2; + float e,w,n,s,u,d,e1,w1,n1,s1,u1,d1; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i >= 0) && (i < N) && (j >= 0) && (j < M) && (k >= 0) && (k < Z)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i+1; if (i1 == N) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == M) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + k1 = k+1; if (k1 == Z) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + + e = Output[(N*M)*k + i1 + N*j]; + w = Output[(N*M)*k + i2 + N*j]; + n = Output[(N*M)*k + i + N*j1]; + s = Output[(N*M)*k + i + N*j2]; + u = Output[(N*M)*k1 + i + N*j]; + d = Output[(N*M)*k2 + i + N*j]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + u1 = u - Output[index]; + d1 = d - Output[index]; + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1 + u1 + d1) - (Output[index] - Input[index])); + } + } + +__global__ void NonLinearDiff3D_kernel(float *Input, float *Output, float lambdaPar, float sigmaPar, float tau, int penaltytype, int N, int M, int Z) + { + int i1,i2,j1,j2,k1,k2; + float e,w,n,s,u,d,e1,w1,n1,s1,u1,d1; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i >= 0) && (i < N) && (j >= 0) && (j < M) && (k >= 0) && (k < Z)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i+1; if (i1 == N) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == M) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + k1 = k+1; if (k1 == Z) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + + e = Output[(N*M)*k + i1 + N*j]; + w = Output[(N*M)*k + i2 + N*j]; + n = Output[(N*M)*k + i + N*j1]; + s = Output[(N*M)*k + i + N*j2]; + u = Output[(N*M)*k1 + i + N*j]; + d = Output[(N*M)*k2 + i + N*j]; + + e1 = e - Output[index]; + w1 = w - Output[index]; + n1 = n - Output[index]; + s1 = s - Output[index]; + u1 = u - Output[index]; + d1 = d - Output[index]; + + + if (penaltytype == 1){ + /* Huber penalty */ + if (abs(e1) > sigmaPar) e1 = signNDF(e1); + else e1 = e1/sigmaPar; + + if (abs(w1) > sigmaPar) w1 = signNDF(w1); + else w1 = w1/sigmaPar; + + if (abs(n1) > sigmaPar) n1 = signNDF(n1); + else n1 = n1/sigmaPar; + + if (abs(s1) > sigmaPar) s1 = signNDF(s1); + else s1 = s1/sigmaPar; + + if (abs(u1) > sigmaPar) u1 = signNDF(u1); + else u1 = u1/sigmaPar; + + if (abs(d1) > sigmaPar) d1 = signNDF(d1); + else d1 = d1/sigmaPar; + } + else if (penaltytype == 2) { + /* Perona-Malik */ + e1 = (e1)/(1.0f + pow((e1/sigmaPar),2)); + w1 = (w1)/(1.0f + pow((w1/sigmaPar),2)); + n1 = (n1)/(1.0f + pow((n1/sigmaPar),2)); + s1 = (s1)/(1.0f + pow((s1/sigmaPar),2)); + u1 = (u1)/(1.0f + pow((u1/sigmaPar),2)); + d1 = (d1)/(1.0f + pow((d1/sigmaPar),2)); + } + else if (penaltytype == 3) { + /* Tukey Biweight */ + if (abs(e1) <= sigmaPar) e1 = e1*pow((1.0f - pow((e1/sigmaPar),2)), 2); + else e1 = 0.0f; + if (abs(w1) <= sigmaPar) w1 = w1*pow((1.0f - pow((w1/sigmaPar),2)), 2); + else w1 = 0.0f; + if (abs(n1) <= sigmaPar) n1 = n1*pow((1.0f - pow((n1/sigmaPar),2)), 2); + else n1 = 0.0f; + if (abs(s1) <= sigmaPar) s1 = s1*pow((1.0f - pow((s1/sigmaPar),2)), 2); + else s1 = 0.0f; + if (abs(u1) <= sigmaPar) u1 = u1*pow((1.0f - pow((u1/sigmaPar),2)), 2); + else u1 = 0.0f; + if (abs(d1) <= sigmaPar) d1 = d1*pow((1.0f - pow((d1/sigmaPar),2)), 2); + else d1 = 0.0f; + } + else printf("%s \n", "No penalty function selected! Use 1,2 or 3."); + + Output[index] += tau*(lambdaPar*(e1 + w1 + n1 + s1 + u1 + d1) - (Output[index] - Input[index])); + } + } + +///////////////////////////////////////////////// +// HOST FUNCTION +extern "C" int NonlDiff_GPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int penaltytype, int N, int M, int Z) +{ + // set up device + int dev = 0; + CHECK(cudaSetDevice(dev)); + float *d_input, *d_output; + float sigmaPar2; + sigmaPar2 = sigmaPar/sqrt(2.0f); + + CHECK(cudaMalloc((void**)&d_input,N*M*Z*sizeof(float))); + CHECK(cudaMalloc((void**)&d_output,N*M*Z*sizeof(float))); + + CHECK(cudaMemcpy(d_input,Input,N*M*Z*sizeof(float),cudaMemcpyHostToDevice)); + CHECK(cudaMemcpy(d_output,Input,N*M*Z*sizeof(float),cudaMemcpyHostToDevice)); + + if (Z == 1) { + /*2D case */ + + dim3 dimBlock(BLKXSIZE2D,BLKYSIZE2D); + dim3 dimGrid(idivup(N,BLKXSIZE2D), idivup(M,BLKYSIZE2D)); + + for(int n=0; n < iterationsNumb; n++) { + if (sigmaPar == 0.0f) { + /* linear diffusion (heat equation) */ + LinearDiff2D_kernel<<<dimGrid,dimBlock>>>(d_input, d_output, lambdaPar, tau, N, M); + CHECK(cudaDeviceSynchronize()); + } + else { + /* nonlinear diffusion */ + NonLinearDiff2D_kernel<<<dimGrid,dimBlock>>>(d_input, d_output, lambdaPar, sigmaPar2, tau, penaltytype, N, M); + CHECK(cudaDeviceSynchronize()); + } + } + } + else { + /*3D case*/ + dim3 dimBlock(BLKXSIZE,BLKYSIZE,BLKZSIZE); + dim3 dimGrid(idivup(N,BLKXSIZE), idivup(M,BLKYSIZE),idivup(Z,BLKZSIZE)); + for(int n=0; n < iterationsNumb; n++) { + if (sigmaPar == 0.0f) { + /* linear diffusion (heat equation) */ + LinearDiff3D_kernel<<<dimGrid,dimBlock>>>(d_input, d_output, lambdaPar, tau, N, M, Z); + CHECK(cudaDeviceSynchronize()); + } + else { + /* nonlinear diffusion */ + NonLinearDiff3D_kernel<<<dimGrid,dimBlock>>>(d_input, d_output, lambdaPar, sigmaPar2, tau, penaltytype, N, M, Z); + CHECK(cudaDeviceSynchronize()); + } + } + + } + CHECK(cudaMemcpy(Output,d_output,N*M*Z*sizeof(float),cudaMemcpyDeviceToHost)); + CHECK(cudaFree(d_input)); + CHECK(cudaFree(d_output)); + //cudaDeviceReset(); + return 0; +} diff --git a/src/Core/regularisers_GPU/NonlDiff_GPU_core.h b/src/Core/regularisers_GPU/NonlDiff_GPU_core.h new file mode 100644 index 0000000..5fe457e --- /dev/null +++ b/src/Core/regularisers_GPU/NonlDiff_GPU_core.h @@ -0,0 +1,8 @@ +#ifndef __NonlDiffGPU_H__ +#define __NonlDiffGPU_H__ +#include "CCPiDefines.h" +#include <stdio.h> + +extern "C" CCPI_EXPORT int NonlDiff_GPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int penaltytype, int N, int M, int Z); + +#endif diff --git a/src/Core/regularisers_GPU/PatchSelect_GPU_core.cu b/src/Core/regularisers_GPU/PatchSelect_GPU_core.cu new file mode 100644 index 0000000..98c8488 --- /dev/null +++ b/src/Core/regularisers_GPU/PatchSelect_GPU_core.cu @@ -0,0 +1,460 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC and Diamond Light Source Ltd. + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * Copyright 2018 Diamond Light Source Ltd. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "PatchSelect_GPU_core.h" +#include "shared.h" + +/* CUDA implementation of non-local weight pre-calculation for non-local priors + * Weights and associated indices are stored into pre-allocated arrays and passed + * to the regulariser + * + * + * Input Parameters: + * 1. 2D grayscale image (classical 3D version will not be supported but rather 2D + dim extension (TODO)) + * 2. Searching window (half-size of the main bigger searching window, e.g. 11) + * 3. Similarity window (half-size of the patch window, e.g. 2) + * 4. The number of neighbours to take (the most prominent after sorting neighbours will be taken) + * 5. noise-related parameter to calculate non-local weights + * + * Output [2D]: + * 1. AR_i - indeces of i neighbours + * 2. AR_j - indeces of j neighbours + * 3. Weights_ij - associated weights + */ + + +#define BLKXSIZE 16 +#define BLKYSIZE 16 +#define idivup(a, b) ( ((a)%(b) != 0) ? (a)/(b)+1 : (a)/(b) ) +#define M_PI 3.14159265358979323846 +#define EPS 1.0e-8 +#define CONSTVECSIZE5 121 +#define CONSTVECSIZE7 225 +#define CONSTVECSIZE9 361 +#define CONSTVECSIZE11 529 +#define CONSTVECSIZE13 729 + +__device__ void swap(float *xp, float *yp) +{ + float temp = *xp; + *xp = *yp; + *yp = temp; +} +__device__ void swapUS(unsigned short *xp, unsigned short *yp) +{ + unsigned short temp = *xp; + *xp = *yp; + *yp = temp; +} + +/********************************************************************************/ +__global__ void IndexSelect2D_5_kernel(float *Ad, unsigned short *H_i_d, unsigned short *H_j_d, float *Weights_d, float *Eucl_Vec_d, int N, int M, int SearchWindow, int SearchW_full, int SimilarWin, int NumNeighb, float h2) +{ + + long i1, j1, i_m, j_m, i_c, j_c, i2, j2, i3, j3, counter, x, y, counterG, index2; + float normsum; + + float Weight_Vec[CONSTVECSIZE5]; + unsigned short ind_i[CONSTVECSIZE5]; + unsigned short ind_j[CONSTVECSIZE5]; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + long index = i*M+j; + + counter = 0; + for(i_m=-SearchWindow; i_m<=SearchWindow; i_m++) { + for(j_m=-SearchWindow; j_m<=SearchWindow; j_m++) { + i1 = i+i_m; + j1 = j+j_m; + if (((i1 >= 0) && (i1 < N)) && ((j1 >= 0) && (j1 < M))) { + normsum = 0.0f; counterG = 0; + for(i_c=-SimilarWin; i_c<=SimilarWin; i_c++) { + for(j_c=-SimilarWin; j_c<=SimilarWin; j_c++) { + i2 = i1 + i_c; + j2 = j1 + j_c; + i3 = i + i_c; + j3 = j + j_c; + if (((i2 >= 0) && (i2 < N)) && ((j2 >= 0) && (j2 < M))) { + if (((i3 >= 0) && (i3 < N)) && ((j3 >= 0) && (j3 < M))) { + normsum += Eucl_Vec_d[counterG]*powf(Ad[i3*M + j3] - Ad[i2*M + j2], 2); + counterG++; + }} + }} + /* writing temporarily into vectors */ + if (normsum > EPS) { + Weight_Vec[counter] = __expf(-normsum/h2); + ind_i[counter] = i1; + ind_j[counter] = j1; + counter++; + } + } + }} + + /* do sorting to choose the most prominent weights [HIGH to LOW] */ + /* and re-arrange indeces accordingly */ + for (x = 0; x < counter-1; x++) { + for (y = 0; y < counter-x-1; y++) { + if (Weight_Vec[y] < Weight_Vec[y+1]) { + swap(&Weight_Vec[y], &Weight_Vec[y+1]); + swapUS(&ind_i[y], &ind_i[y+1]); + swapUS(&ind_j[y], &ind_j[y+1]); + } + } + } + /*sorting loop finished*/ + /*now select the NumNeighb more prominent weights and store into arrays */ + for(x=0; x < NumNeighb; x++) { + index2 = (N*M*x) + index; + H_i_d[index2] = ind_i[x]; + H_j_d[index2] = ind_j[x]; + Weights_d[index2] = Weight_Vec[x]; + } +} +/********************************************************************************/ +__global__ void IndexSelect2D_7_kernel(float *Ad, unsigned short *H_i_d, unsigned short *H_j_d, float *Weights_d, float *Eucl_Vec_d, int N, int M, int SearchWindow, int SearchW_full, int SimilarWin, int NumNeighb, float h2) +{ + + long i1, j1, i_m, j_m, i_c, j_c, i2, j2, i3, j3, counter, x, y, counterG, index2; + float normsum; + + float Weight_Vec[CONSTVECSIZE7]; + unsigned short ind_i[CONSTVECSIZE7]; + unsigned short ind_j[CONSTVECSIZE7]; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + long index = i*M+j; + + counter = 0; + for(i_m=-SearchWindow; i_m<=SearchWindow; i_m++) { + for(j_m=-SearchWindow; j_m<=SearchWindow; j_m++) { + i1 = i+i_m; + j1 = j+j_m; + if (((i1 >= 0) && (i1 < N)) && ((j1 >= 0) && (j1 < M))) { + normsum = 0.0f; counterG = 0; + for(i_c=-SimilarWin; i_c<=SimilarWin; i_c++) { + for(j_c=-SimilarWin; j_c<=SimilarWin; j_c++) { + i2 = i1 + i_c; + j2 = j1 + j_c; + i3 = i + i_c; + j3 = j + j_c; + if (((i2 >= 0) && (i2 < N)) && ((j2 >= 0) && (j2 < M))) { + if (((i3 >= 0) && (i3 < N)) && ((j3 >= 0) && (j3 < M))) { + normsum += Eucl_Vec_d[counterG]*powf(Ad[i3*M + j3] - Ad[i2*M + j2], 2); + counterG++; + }} + }} + /* writing temporarily into vectors */ + if (normsum > EPS) { + Weight_Vec[counter] = __expf(-normsum/h2); + ind_i[counter] = i1; + ind_j[counter] = j1; + counter++; + } + } + }} + + /* do sorting to choose the most prominent weights [HIGH to LOW] */ + /* and re-arrange indeces accordingly */ + for (x = 0; x < counter-1; x++) { + for (y = 0; y < counter-x-1; y++) { + if (Weight_Vec[y] < Weight_Vec[y+1]) { + swap(&Weight_Vec[y], &Weight_Vec[y+1]); + swapUS(&ind_i[y], &ind_i[y+1]); + swapUS(&ind_j[y], &ind_j[y+1]); + } + } + } + /*sorting loop finished*/ + /*now select the NumNeighb more prominent weights and store into arrays */ + for(x=0; x < NumNeighb; x++) { + index2 = (N*M*x) + index; + H_i_d[index2] = ind_i[x]; + H_j_d[index2] = ind_j[x]; + Weights_d[index2] = Weight_Vec[x]; + } +} +__global__ void IndexSelect2D_9_kernel(float *Ad, unsigned short *H_i_d, unsigned short *H_j_d, float *Weights_d, float *Eucl_Vec_d, int N, int M, int SearchWindow, int SearchW_full, int SimilarWin, int NumNeighb, float h2) +{ + + long i1, j1, i_m, j_m, i_c, j_c, i2, j2, i3, j3, counter, x, y, counterG, index2; + float normsum; + + float Weight_Vec[CONSTVECSIZE9]; + unsigned short ind_i[CONSTVECSIZE9]; + unsigned short ind_j[CONSTVECSIZE9]; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + long index = i*M+j; + + counter = 0; + for(i_m=-SearchWindow; i_m<=SearchWindow; i_m++) { + for(j_m=-SearchWindow; j_m<=SearchWindow; j_m++) { + i1 = i+i_m; + j1 = j+j_m; + if (((i1 >= 0) && (i1 < N)) && ((j1 >= 0) && (j1 < M))) { + normsum = 0.0f; counterG = 0; + for(i_c=-SimilarWin; i_c<=SimilarWin; i_c++) { + for(j_c=-SimilarWin; j_c<=SimilarWin; j_c++) { + i2 = i1 + i_c; + j2 = j1 + j_c; + i3 = i + i_c; + j3 = j + j_c; + if (((i2 >= 0) && (i2 < N)) && ((j2 >= 0) && (j2 < M))) { + if (((i3 >= 0) && (i3 < N)) && ((j3 >= 0) && (j3 < M))) { + normsum += Eucl_Vec_d[counterG]*powf(Ad[i3*M + j3] - Ad[i2*M + j2], 2); + counterG++; + }} + }} + /* writing temporarily into vectors */ + if (normsum > EPS) { + Weight_Vec[counter] = expf(-normsum/h2); + ind_i[counter] = i1; + ind_j[counter] = j1; + counter++; + } + } + }} + + /* do sorting to choose the most prominent weights [HIGH to LOW] */ + /* and re-arrange indeces accordingly */ + for (x = 0; x < counter-1; x++) { + for (y = 0; y < counter-x-1; y++) { + if (Weight_Vec[y] < Weight_Vec[y+1]) { + swap(&Weight_Vec[y], &Weight_Vec[y+1]); + swapUS(&ind_i[y], &ind_i[y+1]); + swapUS(&ind_j[y], &ind_j[y+1]); + } + } + } + /*sorting loop finished*/ + /*now select the NumNeighb more prominent weights and store into arrays */ + for(x=0; x < NumNeighb; x++) { + index2 = (N*M*x) + index; + H_i_d[index2] = ind_i[x]; + H_j_d[index2] = ind_j[x]; + Weights_d[index2] = Weight_Vec[x]; + } +} +__global__ void IndexSelect2D_11_kernel(float *Ad, unsigned short *H_i_d, unsigned short *H_j_d, float *Weights_d, float *Eucl_Vec_d, int N, int M, int SearchWindow, int SearchW_full, int SimilarWin, int NumNeighb, float h2) +{ + + long i1, j1, i_m, j_m, i_c, j_c, i2, j2, i3, j3, counter, x, y, counterG, index2; + float normsum; + + float Weight_Vec[CONSTVECSIZE11]; + unsigned short ind_i[CONSTVECSIZE11]; + unsigned short ind_j[CONSTVECSIZE11]; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + long index = i*M+j; + + counter = 0; + for(i_m=-SearchWindow; i_m<=SearchWindow; i_m++) { + for(j_m=-SearchWindow; j_m<=SearchWindow; j_m++) { + i1 = i+i_m; + j1 = j+j_m; + if (((i1 >= 0) && (i1 < N)) && ((j1 >= 0) && (j1 < M))) { + normsum = 0.0f; counterG = 0; + for(i_c=-SimilarWin; i_c<=SimilarWin; i_c++) { + for(j_c=-SimilarWin; j_c<=SimilarWin; j_c++) { + i2 = i1 + i_c; + j2 = j1 + j_c; + i3 = i + i_c; + j3 = j + j_c; + if (((i2 >= 0) && (i2 < N)) && ((j2 >= 0) && (j2 < M))) { + if (((i3 >= 0) && (i3 < N)) && ((j3 >= 0) && (j3 < M))) { + normsum += Eucl_Vec_d[counterG]*powf(Ad[i3*M + j3] - Ad[i2*M + j2], 2); + counterG++; + }} + }} + /* writing temporarily into vectors */ + if (normsum > EPS) { + Weight_Vec[counter] = __expf(-normsum/h2); + ind_i[counter] = i1; + ind_j[counter] = j1; + counter++; + } + } + }} + + /* do sorting to choose the most prominent weights [HIGH to LOW] */ + /* and re-arrange indeces accordingly */ + for (x = 0; x < counter-1; x++) { + for (y = 0; y < counter-x-1; y++) { + if (Weight_Vec[y] < Weight_Vec[y+1]) { + swap(&Weight_Vec[y], &Weight_Vec[y+1]); + swapUS(&ind_i[y], &ind_i[y+1]); + swapUS(&ind_j[y], &ind_j[y+1]); + } + } + } + /*sorting loop finished*/ + /*now select the NumNeighb more prominent weights and store into arrays */ + for(x=0; x < NumNeighb; x++) { + index2 = (N*M*x) + index; + H_i_d[index2] = ind_i[x]; + H_j_d[index2] = ind_j[x]; + Weights_d[index2] = Weight_Vec[x]; + } +} +__global__ void IndexSelect2D_13_kernel(float *Ad, unsigned short *H_i_d, unsigned short *H_j_d, float *Weights_d, float *Eucl_Vec_d, int N, int M, int SearchWindow, int SearchW_full, int SimilarWin, int NumNeighb, float h2) +{ + + long i1, j1, i_m, j_m, i_c, j_c, i2, j2, i3, j3, counter, x, y, counterG, index2; + float normsum; + + float Weight_Vec[CONSTVECSIZE13]; + unsigned short ind_i[CONSTVECSIZE13]; + unsigned short ind_j[CONSTVECSIZE13]; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + long index = i*M+j; + + counter = 0; + for(i_m=-SearchWindow; i_m<=SearchWindow; i_m++) { + for(j_m=-SearchWindow; j_m<=SearchWindow; j_m++) { + i1 = i+i_m; + j1 = j+j_m; + if (((i1 >= 0) && (i1 < N)) && ((j1 >= 0) && (j1 < M))) { + normsum = 0.0f; counterG = 0; + for(i_c=-SimilarWin; i_c<=SimilarWin; i_c++) { + for(j_c=-SimilarWin; j_c<=SimilarWin; j_c++) { + i2 = i1 + i_c; + j2 = j1 + j_c; + i3 = i + i_c; + j3 = j + j_c; + if (((i2 >= 0) && (i2 < N)) && ((j2 >= 0) && (j2 < M))) { + if (((i3 >= 0) && (i3 < N)) && ((j3 >= 0) && (j3 < M))) { + normsum += Eucl_Vec_d[counterG]*powf(Ad[i3*M + j3] - Ad[i2*M + j2], 2); + counterG++; + }} + }} + /* writing temporarily into vectors */ + if (normsum > EPS) { + Weight_Vec[counter] = __expf(-normsum/h2); + ind_i[counter] = i1; + ind_j[counter] = j1; + counter++; + } + } + }} + + /* do sorting to choose the most prominent weights [HIGH to LOW] */ + /* and re-arrange indeces accordingly */ + for (x = 0; x < counter-1; x++) { + for (y = 0; y < counter-x-1; y++) { + if (Weight_Vec[y] < Weight_Vec[y+1]) { + swap(&Weight_Vec[y], &Weight_Vec[y+1]); + swapUS(&ind_i[y], &ind_i[y+1]); + swapUS(&ind_j[y], &ind_j[y+1]); + } + } + } + /*sorting loop finished*/ + /*now select the NumNeighb more prominent weights and store into arrays */ + for(x=0; x < NumNeighb; x++) { + index2 = (N*M*x) + index; + H_i_d[index2] = ind_i[x]; + H_j_d[index2] = ind_j[x]; + Weights_d[index2] = Weight_Vec[x]; + } +} + + +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ +/********************* MAIN HOST FUNCTION ******************/ +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ +extern "C" int PatchSelect_GPU_main(float *A, unsigned short *H_i, unsigned short *H_j, float *Weights, int N, int M, int SearchWindow, int SimilarWin, int NumNeighb, float h) +{ + int deviceCount = -1; // number of devices + cudaGetDeviceCount(&deviceCount); + if (deviceCount == 0) { + fprintf(stderr, "No CUDA devices found\n"); + return -1; + } + + int SearchW_full, SimilW_full, counterG, i, j; + float *Ad, *Weights_d, h2, *Eucl_Vec, *Eucl_Vec_d; + unsigned short *H_i_d, *H_j_d; + h2 = h*h; + + dim3 dimBlock(BLKXSIZE,BLKYSIZE); + dim3 dimGrid(idivup(N,BLKXSIZE), idivup(M,BLKYSIZE)); + + SearchW_full = (2*SearchWindow + 1)*(2*SearchWindow + 1); /* the full searching window size */ + SimilW_full = (2*SimilarWin + 1)*(2*SimilarWin + 1); /* the full similarity window size */ + + /* generate a 2D Gaussian kernel for NLM procedure */ + Eucl_Vec = (float*) calloc (SimilW_full,sizeof(float)); + counterG = 0; + for(i=-SimilarWin; i<=SimilarWin; i++) { + for(j=-SimilarWin; j<=SimilarWin; j++) { + Eucl_Vec[counterG] = (float)exp(-(pow(((float) i), 2) + pow(((float) j), 2))/(2.0*SimilarWin*SimilarWin)); + counterG++; + }} /*main neighb loop */ + + + /*allocate space on the device*/ + checkCudaErrors( cudaMalloc((void**)&Ad, N*M*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&H_i_d, N*M*NumNeighb*sizeof(unsigned short)) ); + checkCudaErrors( cudaMalloc((void**)&H_j_d, N*M*NumNeighb*sizeof(unsigned short)) ); + checkCudaErrors( cudaMalloc((void**)&Weights_d, N*M*NumNeighb*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&Eucl_Vec_d, SimilW_full*sizeof(float)) ); + + /* copy data from the host to the device */ + checkCudaErrors( cudaMemcpy(Ad,A,N*M*sizeof(float),cudaMemcpyHostToDevice) ); + checkCudaErrors( cudaMemcpy(Eucl_Vec_d,Eucl_Vec,SimilW_full*sizeof(float),cudaMemcpyHostToDevice) ); + + /********************** Run CUDA kernel here ********************/ + if (SearchWindow == 5) IndexSelect2D_5_kernel<<<dimGrid,dimBlock>>>(Ad, H_i_d, H_j_d, Weights_d, Eucl_Vec_d, N, M, SearchWindow, SearchW_full, SimilarWin, NumNeighb, h2); + else if (SearchWindow == 7) IndexSelect2D_7_kernel<<<dimGrid,dimBlock>>>(Ad, H_i_d, H_j_d, Weights_d, Eucl_Vec_d, N, M, SearchWindow, SearchW_full, SimilarWin, NumNeighb, h2); + else if (SearchWindow == 9) IndexSelect2D_9_kernel<<<dimGrid,dimBlock>>>(Ad, H_i_d, H_j_d, Weights_d, Eucl_Vec_d, N, M, SearchWindow, SearchW_full, SimilarWin, NumNeighb, h2); + else if (SearchWindow == 11) IndexSelect2D_11_kernel<<<dimGrid,dimBlock>>>(Ad, H_i_d, H_j_d, Weights_d, Eucl_Vec_d, N, M, SearchWindow, SearchW_full, SimilarWin, NumNeighb, h2); + else if (SearchWindow == 13) IndexSelect2D_13_kernel<<<dimGrid,dimBlock>>>(Ad, H_i_d, H_j_d, Weights_d, Eucl_Vec_d, N, M, SearchWindow, SearchW_full, SimilarWin, NumNeighb, h2); + else { + fprintf(stderr, "Select the searching window size from 5, 7, 9, 11 or 13\n"); + return -1;} + checkCudaErrors(cudaPeekAtLastError() ); + checkCudaErrors(cudaDeviceSynchronize()); + /***************************************************************/ + + checkCudaErrors(cudaMemcpy(H_i, H_i_d, N*M*NumNeighb*sizeof(unsigned short),cudaMemcpyDeviceToHost) ); + checkCudaErrors(cudaMemcpy(H_j, H_j_d, N*M*NumNeighb*sizeof(unsigned short),cudaMemcpyDeviceToHost) ); + checkCudaErrors(cudaMemcpy(Weights, Weights_d, N*M*NumNeighb*sizeof(float),cudaMemcpyDeviceToHost) ); + + + cudaFree(Ad); + cudaFree(H_i_d); + cudaFree(H_j_d); + cudaFree(Weights_d); + cudaFree(Eucl_Vec_d); + cudaDeviceReset(); + return 0; +} diff --git a/src/Core/regularisers_GPU/PatchSelect_GPU_core.h b/src/Core/regularisers_GPU/PatchSelect_GPU_core.h new file mode 100644 index 0000000..8c124d3 --- /dev/null +++ b/src/Core/regularisers_GPU/PatchSelect_GPU_core.h @@ -0,0 +1,8 @@ +#ifndef __NLREG_KERNELS_H_ +#define __NLREG_KERNELS_H_ +#include "CCPiDefines.h" +#include <stdio.h> + +extern "C" CCPI_EXPORT int PatchSelect_GPU_main(float *A, unsigned short *H_i, unsigned short *H_j, float *Weights, int N, int M, int SearchWindow, int SimilarWin, int NumNeighb, float h); + +#endif diff --git a/src/Core/regularisers_GPU/TGV_GPU_core.cu b/src/Core/regularisers_GPU/TGV_GPU_core.cu new file mode 100644 index 0000000..58b2c41 --- /dev/null +++ b/src/Core/regularisers_GPU/TGV_GPU_core.cu @@ -0,0 +1,625 @@ + /* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "TGV_GPU_core.h" +#include "shared.h" + +/* CUDA implementation of Primal-Dual denoising method for + * Total Generilized Variation (TGV)-L2 model [1] (2D/3D case) + * + * Input Parameters: + * 1. Noisy image/volume (2D/3D) + * 2. lambda - regularisation parameter + * 3. parameter to control the first-order term (alpha1) + * 4. parameter to control the second-order term (alpha0) + * 5. Number of Chambolle-Pock (Primal-Dual) iterations + * 6. Lipshitz constant (default is 12) + * + * Output: + * Filtered/regulariaed image + * + * References: + * [1] K. Bredies "Total Generalized Variation" + */ + +#define BLKXSIZE 8 +#define BLKYSIZE 8 +#define BLKZSIZE 8 + +#define BLKXSIZE2D 16 +#define BLKYSIZE2D 16 +#define EPS 1.0e-7 +#define idivup(a, b) ( ((a)%(b) != 0) ? (a)/(b)+1 : (a)/(b) ) + + +/********************************************************************/ +/***************************2D Functions*****************************/ +/********************************************************************/ +__global__ void DualP_2D_kernel(float *U, float *V1, float *V2, float *P1, float *P2, int dimX, int dimY, float sigma) +{ + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + /* symmetric boundary conditions (Neuman) */ + if (i == dimX-1) P1[index] += sigma*((U[j*dimX+(i-1)] - U[index]) - V1[index]); + else P1[index] += sigma*((U[j*dimX+(i+1)] - U[index]) - V1[index]); + if (j == dimY-1) P2[index] += sigma*((U[(j-1)*dimX+i] - U[index]) - V2[index]); + else P2[index] += sigma*((U[(j+1)*dimX+i] - U[index]) - V2[index]); + } + return; +} + +__global__ void ProjP_2D_kernel(float *P1, float *P2, int dimX, int dimY, float alpha1) +{ + float grad_magn; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + + grad_magn = sqrt(pow(P1[index],2) + pow(P2[index],2)); + grad_magn = grad_magn/alpha1; + if (grad_magn > 1.0f) { + P1[index] /= grad_magn; + P2[index] /= grad_magn; + } + } + return; +} + +__global__ void DualQ_2D_kernel(float *V1, float *V2, float *Q1, float *Q2, float *Q3, int dimX, int dimY, float sigma) +{ + float q1, q2, q11, q22; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + /* symmetric boundary conditions (Neuman) */ + q1 = 0.0f; q11 = 0.0f; q2 = 0.0f; q22 = 0.0f; + /* boundary conditions (Neuman) */ + if (i != dimX-1){ + q1 = V1[j*dimX+(i+1)] - V1[index]; + q11 = V2[j*dimX+(i+1)] - V2[index]; + } + if (j != dimY-1) { + q2 = V2[(j+1)*dimX+i] - V2[index]; + q22 = V1[(j+1)*dimX+i] - V1[index]; + } + Q1[index] += sigma*(q1); + Q2[index] += sigma*(q2); + Q3[index] += sigma*(0.5f*(q11 + q22)); + } + return; +} + +__global__ void ProjQ_2D_kernel(float *Q1, float *Q2, float *Q3, int dimX, int dimY, float alpha0) +{ + float grad_magn; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + grad_magn = sqrt(pow(Q1[index],2) + pow(Q2[index],2) + 2*pow(Q3[index],2)); + grad_magn = grad_magn/alpha0; + if (grad_magn > 1.0f) { + Q1[index] /= grad_magn; + Q2[index] /= grad_magn; + Q3[index] /= grad_magn; + } + } + return; +} + +__global__ void DivProjP_2D_kernel(float *U, float *U0, float *P1, float *P2, int dimX, int dimY, float lambda, float tau) +{ + float P_v1, P_v2, div; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + + if (i == 0) P_v1 = P1[index]; + else P_v1 = P1[index] - P1[j*dimX+(i-1)]; + if (j == 0) P_v2 = P2[index]; + else P_v2 = P2[index] - P2[(j-1)*dimX+i]; + div = P_v1 + P_v2; + U[index] = (lambda*(U[index] + tau*div) + tau*U0[index])/(lambda + tau); + } + return; +} + +__global__ void UpdV_2D_kernel(float *V1, float *V2, float *P1, float *P2, float *Q1, float *Q2, float *Q3, int dimX, int dimY, float tau) +{ + float q1, q3_x, q2, q3_y, div1, div2; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + dimX*j; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY)) { + q2 = 0.0f; q3_y = 0.0f; q1 = 0.0f; q3_x = 0.0; + /* boundary conditions (Neuman) */ + if (i != 0) { + q1 = Q1[index] - Q1[j*dimX+(i-1)]; + q3_x = Q3[index] - Q3[j*dimX+(i-1)]; + } + if (j != 0) { + q2 = Q2[index] - Q2[(j-1)*dimX+i]; + q3_y = Q3[index] - Q3[(j-1)*dimX+i]; + } + div1 = q1 + q3_y; + div2 = q3_x + q2; + V1[index] += tau*(P1[index] + div1); + V2[index] += tau*(P2[index] + div2); + } + return; +} + +__global__ void copyIm_TGV_kernel(float *U, float *U_old, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + U_old[index] = U[index]; + } +} + +__global__ void copyIm_TGV_kernel_ar2(float *V1, float *V2, float *V1_old, float *V2_old, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + V1_old[index] = V1[index]; + V2_old[index] = V2[index]; + } +} + +__global__ void newU_kernel(float *U, float *U_old, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + U[index] = 2.0f*U[index] - U_old[index]; + } +} + + +__global__ void newU_kernel_ar2(float *V1, float *V2, float *V1_old, float *V2_old, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + V1[index] = 2.0f*V1[index] - V1_old[index]; + V2[index] = 2.0f*V2[index] - V2_old[index]; + } +} +/********************************************************************/ +/***************************3D Functions*****************************/ +/********************************************************************/ +__global__ void DualP_3D_kernel(float *U, float *V1, float *V2, float *V3, float *P1, float *P2, float *P3, int dimX, int dimY, int dimZ, float sigma) +{ + int index; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + index = (dimX*dimY)*k + j*dimX+i; + /* symmetric boundary conditions (Neuman) */ + if (i == dimX-1) P1[index] += sigma*((U[(dimX*dimY)*k + j*dimX+(i-1)] - U[index]) - V1[index]); + else P1[index] += sigma*((U[(dimX*dimY)*k + j*dimX+(i+1)] - U[index]) - V1[index]); + if (j == dimY-1) P2[index] += sigma*((U[(dimX*dimY)*k + (j-1)*dimX+i] - U[index]) - V2[index]); + else P2[index] += sigma*((U[(dimX*dimY)*k + (j+1)*dimX+i] - U[index]) - V2[index]); + if (k == dimZ-1) P3[index] += sigma*((U[(dimX*dimY)*(k-1) + j*dimX+i] - U[index]) - V3[index]); + else P3[index] += sigma*((U[(dimX*dimY)*(k+1) + j*dimX+i] - U[index]) - V3[index]); + } + return; +} + +__global__ void ProjP_3D_kernel(float *P1, float *P2, float *P3, int dimX, int dimY, int dimZ, float alpha1) +{ + float grad_magn; + int index; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + index = (dimX*dimY)*k + j*dimX+i; + + grad_magn = (sqrtf(pow(P1[index],2) + pow(P2[index],2) + pow(P3[index],2)))/alpha1; + if (grad_magn > 1.0f) { + P1[index] /= grad_magn; + P2[index] /= grad_magn; + P3[index] /= grad_magn; + } + } + return; +} + +__global__ void DualQ_3D_kernel(float *V1, float *V2, float *V3, float *Q1, float *Q2, float *Q3, float *Q4, float *Q5, float *Q6, int dimX, int dimY, int dimZ, float sigma) +{ + int index; + float q1, q2, q3, q11, q22, q33, q44, q55, q66; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + index = (dimX*dimY)*k + j*dimX+i; + q1 = 0.0f; q11 = 0.0f; q33 = 0.0f; q2 = 0.0f; q22 = 0.0f; q55 = 0.0f; q3 = 0.0f; q44 = 0.0f; q66 = 0.0f; + /* symmetric boundary conditions (Neuman) */ + if (i != dimX-1){ + q1 = V1[(dimX*dimY)*k + j*dimX+(i+1)] - V1[index]; + q11 = V2[(dimX*dimY)*k + j*dimX+(i+1)] - V2[index]; + q33 = V3[(dimX*dimY)*k + j*dimX+(i+1)] - V3[index]; + } + if (j != dimY-1) { + q2 = V2[(dimX*dimY)*k + (j+1)*dimX+i] - V2[index]; + q22 = V1[(dimX*dimY)*k + (j+1)*dimX+i] - V1[index]; + q55 = V3[(dimX*dimY)*k + (j+1)*dimX+i] - V3[index]; + } + if (k != dimZ-1) { + q3 = V3[(dimX*dimY)*(k+1) + j*dimX+i] - V3[index]; + q44 = V1[(dimX*dimY)*(k+1) + j*dimX+i] - V1[index]; + q66 = V2[(dimX*dimY)*(k+1) + j*dimX+i] - V2[index]; + } + + Q1[index] += sigma*(q1); /*Q11*/ + Q2[index] += sigma*(q2); /*Q22*/ + Q3[index] += sigma*(q3); /*Q33*/ + Q4[index] += sigma*(0.5f*(q11 + q22)); /* Q21 / Q12 */ + Q5[index] += sigma*(0.5f*(q33 + q44)); /* Q31 / Q13 */ + Q6[index] += sigma*(0.5f*(q55 + q66)); /* Q32 / Q23 */ + } + return; +} + + +__global__ void ProjQ_3D_kernel(float *Q1, float *Q2, float *Q3, float *Q4, float *Q5, float *Q6, int dimX, int dimY, int dimZ, float alpha0) +{ + float grad_magn; + int index; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + index = (dimX*dimY)*k + j*dimX+i; + + grad_magn = sqrtf(pow(Q1[index],2) + pow(Q2[index],2) + pow(Q3[index],2) + 2.0f*pow(Q4[index],2) + 2.0f*pow(Q5[index],2) + 2.0f*pow(Q6[index],2)); + grad_magn = grad_magn/alpha0; + if (grad_magn > 1.0f) { + Q1[index] /= grad_magn; + Q2[index] /= grad_magn; + Q3[index] /= grad_magn; + Q4[index] /= grad_magn; + Q5[index] /= grad_magn; + Q6[index] /= grad_magn; + } + } + return; +} +__global__ void DivProjP_3D_kernel(float *U, float *U0, float *P1, float *P2, float *P3, int dimX, int dimY, int dimZ, float lambda, float tau) +{ + float P_v1, P_v2, P_v3, div; + int index; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + index = (dimX*dimY)*k + j*dimX+i; + + if (i == 0) P_v1 = P1[index]; + else P_v1 = P1[index] - P1[(dimX*dimY)*k + j*dimX+(i-1)]; + if (j == 0) P_v2 = P2[index]; + else P_v2 = P2[index] - P2[(dimX*dimY)*k + (j-1)*dimX+i]; + if (k == 0) P_v3 = P3[index]; + else P_v3 = P3[index] - P3[(dimX*dimY)*(k-1) + (j)*dimX+i]; + + div = P_v1 + P_v2 + P_v3; + U[index] = (lambda*(U[index] + tau*div) + tau*U0[index])/(lambda + tau); + } + return; +} +__global__ void UpdV_3D_kernel(float *V1, float *V2, float *V3, float *P1, float *P2, float *P3, float *Q1, float *Q2, float *Q3, float *Q4, float *Q5, float *Q6, int dimX, int dimY, int dimZ, float tau) +{ + float q1, q4x, q5x, q2, q4y, q6y, q6z, q5z, q3, div1, div2, div3; + int index; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + index = (dimX*dimY)*k + j*dimX+i; + + q1 = 0.0f; q4x= 0.0f; q5x= 0.0f; q2= 0.0f; q4y= 0.0f; q6y= 0.0f; q6z= 0.0f; q5z= 0.0f; q3= 0.0f; + /* Q1 - Q11, Q2 - Q22, Q3 - Q33, Q4 - Q21/Q12, Q5 - Q31/Q13, Q6 - Q32/Q23*/ + /* symmetric boundary conditions (Neuman) */ + if (i != 0) { + q1 = Q1[index] - Q1[(dimX*dimY)*k + j*dimX+(i-1)]; + q4x = Q4[index] - Q4[(dimX*dimY)*k + j*dimX+(i-1)]; + q5x = Q5[index] - Q5[(dimX*dimY)*k + j*dimX+(i-1)]; + } + if (j != 0) { + q2 = Q2[index] - Q2[(dimX*dimY)*k + (j-1)*dimX+i]; + q4y = Q4[index] - Q4[(dimX*dimY)*k + (j-1)*dimX+i]; + q6y = Q6[index] - Q6[(dimX*dimY)*k + (j-1)*dimX+i]; + } + if (k != 0) { + q6z = Q6[index] - Q6[(dimX*dimY)*(k-1) + (j)*dimX+i]; + q5z = Q5[index] - Q5[(dimX*dimY)*(k-1) + (j)*dimX+i]; + q3 = Q3[index] - Q3[(dimX*dimY)*(k-1) + (j)*dimX+i]; + } + div1 = q1 + q4y + q5z; + div2 = q4x + q2 + q6z; + div3 = q5x + q6y + q3; + + V1[index] += tau*(P1[index] + div1); + V2[index] += tau*(P2[index] + div2); + V3[index] += tau*(P3[index] + div3); + } + return; +} + +__global__ void copyIm_TGV_kernel3D(float *U, float *U_old, int dimX, int dimY, int dimZ, int num_total) +{ + int index; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + index = (dimX*dimY)*k + j*dimX+i; + + if (index < num_total) { + U_old[index] = U[index]; + } +} + +__global__ void copyIm_TGV_kernel3D_ar3(float *V1, float *V2, float *V3, float *V1_old, float *V2_old, float *V3_old, int dimX, int dimY, int dimZ, int num_total) +{ + int index; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + index = (dimX*dimY)*k + j*dimX+i; + + if (index < num_total) { + V1_old[index] = V1[index]; + V2_old[index] = V2[index]; + V3_old[index] = V3[index]; + } +} + +__global__ void newU_kernel3D(float *U, float *U_old, int dimX, int dimY, int dimZ, int num_total) +{ + int index; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + index = (dimX*dimY)*k + j*dimX+i; + + if (index < num_total) { + U[index] = 2.0f*U[index] - U_old[index]; + } +} + +__global__ void newU_kernel3D_ar3(float *V1, float *V2, float *V3, float *V1_old, float *V2_old, float *V3_old, int dimX, int dimY, int dimZ, int num_total) +{ + int index; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + index = (dimX*dimY)*k + j*dimX+i; + + if (index < num_total) { + V1[index] = 2.0f*V1[index] - V1_old[index]; + V2[index] = 2.0f*V2[index] - V2_old[index]; + V3[index] = 2.0f*V3[index] - V3_old[index]; + } +} + +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ +/************************ MAIN HOST FUNCTION ***********************/ +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ +extern "C" int TGV_GPU_main(float *U0, float *U, float lambda, float alpha1, float alpha0, int iterationsNumb, float L2, int dimX, int dimY, int dimZ) +{ + int dimTotal, dev = 0; + CHECK(cudaSetDevice(dev)); + + dimTotal = dimX*dimY*dimZ; + + float *U_old, *d_U0, *d_U, *P1, *P2, *Q1, *Q2, *Q3, *V1, *V1_old, *V2, *V2_old, tau, sigma; + tau = pow(L2,-0.5); + sigma = pow(L2,-0.5); + + CHECK(cudaMalloc((void**)&d_U0,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&d_U,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&U_old,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&P1,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&P2,dimTotal*sizeof(float))); + + CHECK(cudaMalloc((void**)&Q1,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&Q2,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&Q3,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&V1,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&V2,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&V1_old,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&V2_old,dimTotal*sizeof(float))); + + CHECK(cudaMemcpy(d_U0,U0,dimTotal*sizeof(float),cudaMemcpyHostToDevice)); + CHECK(cudaMemcpy(d_U,U0,dimTotal*sizeof(float),cudaMemcpyHostToDevice)); + + if (dimZ == 1) { + /*2D case */ + dim3 dimBlock(BLKXSIZE2D,BLKYSIZE2D); + dim3 dimGrid(idivup(dimX,BLKXSIZE2D), idivup(dimY,BLKYSIZE2D)); + + for(int n=0; n < iterationsNumb; n++) { + + /* Calculate Dual Variable P */ + DualP_2D_kernel<<<dimGrid,dimBlock>>>(d_U, V1, V2, P1, P2, dimX, dimY, sigma); + CHECK(cudaDeviceSynchronize()); + /*Projection onto convex set for P*/ + ProjP_2D_kernel<<<dimGrid,dimBlock>>>(P1, P2, dimX, dimY, alpha1); + CHECK(cudaDeviceSynchronize()); + /* Calculate Dual Variable Q */ + DualQ_2D_kernel<<<dimGrid,dimBlock>>>(V1, V2, Q1, Q2, Q3, dimX, dimY, sigma); + CHECK(cudaDeviceSynchronize()); + /*Projection onto convex set for Q*/ + ProjQ_2D_kernel<<<dimGrid,dimBlock>>>(Q1, Q2, Q3, dimX, dimY, alpha0); + CHECK(cudaDeviceSynchronize()); + /*saving U into U_old*/ + copyIm_TGV_kernel<<<dimGrid,dimBlock>>>(d_U, U_old, dimX, dimY, dimTotal); + CHECK(cudaDeviceSynchronize()); + /*adjoint operation -> divergence and projection of P*/ + DivProjP_2D_kernel<<<dimGrid,dimBlock>>>(d_U, d_U0, P1, P2, dimX, dimY, lambda, tau); + CHECK(cudaDeviceSynchronize()); + /*get updated solution U*/ + newU_kernel<<<dimGrid,dimBlock>>>(d_U, U_old, dimX, dimY, dimTotal); + CHECK(cudaDeviceSynchronize()); + /*saving V into V_old*/ + copyIm_TGV_kernel_ar2<<<dimGrid,dimBlock>>>(V1, V2, V1_old, V2_old, dimX, dimY, dimTotal); + CHECK(cudaDeviceSynchronize()); + /* upd V*/ + UpdV_2D_kernel<<<dimGrid,dimBlock>>>(V1, V2, P1, P2, Q1, Q2, Q3, dimX, dimY, tau); + CHECK(cudaDeviceSynchronize()); + /*get new V*/ + newU_kernel_ar2<<<dimGrid,dimBlock>>>(V1, V2, V1_old, V2_old, dimX, dimY, dimTotal); + CHECK(cudaDeviceSynchronize()); + } + } + else { + /*3D case */ + dim3 dimBlock(BLKXSIZE,BLKYSIZE,BLKZSIZE); + dim3 dimGrid(idivup(dimX,BLKXSIZE), idivup(dimY,BLKYSIZE),idivup(dimZ,BLKXSIZE)); + + float *P3, *Q4, *Q5, *Q6, *V3, *V3_old; + + CHECK(cudaMalloc((void**)&P3,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&Q4,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&Q5,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&Q6,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&V3,dimTotal*sizeof(float))); + CHECK(cudaMalloc((void**)&V3_old,dimTotal*sizeof(float))); + + for(int n=0; n < iterationsNumb; n++) { + + /* Calculate Dual Variable P */ + DualP_3D_kernel<<<dimGrid,dimBlock>>>(d_U, V1, V2, V3, P1, P2, P3, dimX, dimY, dimZ, sigma); + CHECK(cudaDeviceSynchronize()); + /*Projection onto convex set for P*/ + ProjP_3D_kernel<<<dimGrid,dimBlock>>>(P1, P2, P3, dimX, dimY, dimZ, alpha1); + CHECK(cudaDeviceSynchronize()); + /* Calculate Dual Variable Q */ + DualQ_3D_kernel<<<dimGrid,dimBlock>>>(V1, V2, V3, Q1, Q2, Q3, Q4, Q5, Q6, dimX, dimY, dimZ, sigma); + CHECK(cudaDeviceSynchronize()); + /*Projection onto convex set for Q*/ + ProjQ_3D_kernel<<<dimGrid,dimBlock>>>(Q1, Q2, Q3, Q4, Q5, Q6, dimX, dimY, dimZ, alpha0); + CHECK(cudaDeviceSynchronize()); + /*saving U into U_old*/ + copyIm_TGV_kernel3D<<<dimGrid,dimBlock>>>(d_U, U_old, dimX, dimY, dimZ, dimTotal); + CHECK(cudaDeviceSynchronize()); + /*adjoint operation -> divergence and projection of P*/ + DivProjP_3D_kernel<<<dimGrid,dimBlock>>>(d_U, d_U0, P1, P2, P3, dimX, dimY, dimZ, lambda, tau); + CHECK(cudaDeviceSynchronize()); + /*get updated solution U*/ + newU_kernel3D<<<dimGrid,dimBlock>>>(d_U, U_old, dimX, dimY, dimZ, dimTotal); + CHECK(cudaDeviceSynchronize()); + /*saving V into V_old*/ + copyIm_TGV_kernel3D_ar3<<<dimGrid,dimBlock>>>(V1, V2, V3, V1_old, V2_old, V3_old, dimX, dimY, dimZ, dimTotal); + CHECK(cudaDeviceSynchronize()); + /* upd V*/ + UpdV_3D_kernel<<<dimGrid,dimBlock>>>(V1, V2, V3, P1, P2, P3, Q1, Q2, Q3, Q4, Q5, Q6, dimX, dimY, dimZ, tau); + CHECK(cudaDeviceSynchronize()); + /*get new V*/ + newU_kernel3D_ar3<<<dimGrid,dimBlock>>>(V1, V2, V3, V1_old, V2_old, V3_old, dimX, dimY, dimZ, dimTotal); + CHECK(cudaDeviceSynchronize()); + } + + CHECK(cudaFree(Q4)); + CHECK(cudaFree(Q5)); + CHECK(cudaFree(Q6)); + CHECK(cudaFree(P3)); + CHECK(cudaFree(V3)); + CHECK(cudaFree(V3_old)); + } + + CHECK(cudaMemcpy(U,d_U,dimTotal*sizeof(float),cudaMemcpyDeviceToHost)); + CHECK(cudaFree(d_U0)); + CHECK(cudaFree(d_U)); + CHECK(cudaFree(U_old)); + CHECK(cudaFree(P1)); + CHECK(cudaFree(P2)); + + CHECK(cudaFree(Q1)); + CHECK(cudaFree(Q2)); + CHECK(cudaFree(Q3)); + CHECK(cudaFree(V1)); + CHECK(cudaFree(V2)); + CHECK(cudaFree(V1_old)); + CHECK(cudaFree(V2_old)); + return 0; +} diff --git a/src/Core/regularisers_GPU/TGV_GPU_core.h b/src/Core/regularisers_GPU/TGV_GPU_core.h new file mode 100644 index 0000000..9f73d1c --- /dev/null +++ b/src/Core/regularisers_GPU/TGV_GPU_core.h @@ -0,0 +1,8 @@ +#ifndef __TGV_GPU_H__ +#define __TGV_GPU_H__ +#include "CCPiDefines.h" +#include <stdio.h> + +extern "C" CCPI_EXPORT int TGV_GPU_main(float *U0, float *U, float lambda, float alpha1, float alpha0, int iterationsNumb, float L2, int dimX, int dimY, int dimZ); + +#endif diff --git a/src/Core/regularisers_GPU/TV_FGP_GPU_core.cu b/src/Core/regularisers_GPU/TV_FGP_GPU_core.cu new file mode 100755 index 0000000..b371c5d --- /dev/null +++ b/src/Core/regularisers_GPU/TV_FGP_GPU_core.cu @@ -0,0 +1,564 @@ + /* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "TV_FGP_GPU_core.h" +#include "shared.h" +#include <thrust/device_vector.h> +#include <thrust/transform_reduce.h> + +/* CUDA implementation of FGP-TV [1] denoising/regularization model (2D/3D case) + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambdaPar - regularization parameter + * 3. Number of iterations + * 4. eplsilon: tolerance constant + * 5. TV-type: methodTV - 'iso' (0) or 'l1' (1) + * 6. nonneg: 'nonnegativity (0 is OFF by default) + * 7. print information: 0 (off) or 1 (on) + * + * Output: + * [1] Filtered/regularized image + * + * This function is based on the Matlab's code and paper by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + */ + + +#define BLKXSIZE2D 16 +#define BLKYSIZE2D 16 + +#define BLKXSIZE 8 +#define BLKYSIZE 8 +#define BLKZSIZE 8 + +#define idivup(a, b) ( ((a)%(b) != 0) ? (a)/(b)+1 : (a)/(b) ) +struct square { __host__ __device__ float operator()(float x) { return x * x; } }; + +/************************************************/ +/*****************2D modules*********************/ +/************************************************/ +__global__ void Obj_func2D_kernel(float *Ad, float *D, float *R1, float *R2, int N, int M, int ImSize, float lambda) +{ + + float val1,val2; + + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + if (xIndex <= 0) {val1 = 0.0f;} else {val1 = R1[(xIndex-1) + N*yIndex];} + if (yIndex <= 0) {val2 = 0.0f;} else {val2 = R2[xIndex + N*(yIndex-1)];} + //Write final result to global memory + D[index] = Ad[index] - lambda*(R1[index] + R2[index] - val1 - val2); + } + return; +} + +__global__ void Grad_func2D_kernel(float *P1, float *P2, float *D, float *R1, float *R2, int N, int M, int ImSize, float multip) +{ + + float val1,val2; + + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + + /* boundary conditions */ + if (xIndex >= N-1) val1 = 0.0f; else val1 = D[index] - D[(xIndex+1) + N*yIndex]; + if (yIndex >= M-1) val2 = 0.0f; else val2 = D[index] - D[(xIndex) + N*(yIndex + 1)]; + + //Write final result to global memory + P1[index] = R1[index] + multip*val1; + P2[index] = R2[index] + multip*val2; + } + return; +} + +__global__ void Proj_func2D_iso_kernel(float *P1, float *P2, int N, int M, int ImSize) +{ + + float denom; + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + denom = pow(P1[index],2) + pow(P2[index],2); + if (denom > 1.0f) { + P1[index] = P1[index]/sqrt(denom); + P2[index] = P2[index]/sqrt(denom); + } + } + return; +} +__global__ void Proj_func2D_aniso_kernel(float *P1, float *P2, int N, int M, int ImSize) +{ + + float val1, val2; + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + val1 = abs(P1[index]); + val2 = abs(P2[index]); + if (val1 < 1.0f) {val1 = 1.0f;} + if (val2 < 1.0f) {val2 = 1.0f;} + P1[index] = P1[index]/val1; + P2[index] = P2[index]/val2; + } + return; +} +__global__ void Rupd_func2D_kernel(float *P1, float *P1_old, float *P2, float *P2_old, float *R1, float *R2, float tkp1, float tk, float multip2, int N, int M, int ImSize) +{ + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + R1[index] = P1[index] + multip2*(P1[index] - P1_old[index]); + R2[index] = P2[index] + multip2*(P2[index] - P2_old[index]); + } + return; +} +__global__ void nonneg2D_kernel(float* Output, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + if (Output[index] < 0.0f) Output[index] = 0.0f; + } +} +/************************************************/ +/*****************3D modules*********************/ +/************************************************/ +__global__ void Obj_func3D_kernel(float *Ad, float *D, float *R1, float *R2, float *R3, int N, int M, int Z, int ImSize, float lambda) +{ + + float val1,val2,val3; + + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + if (i <= 0) {val1 = 0.0f;} else {val1 = R1[(N*M)*(k) + (i-1) + N*j];} + if (j <= 0) {val2 = 0.0f;} else {val2 = R2[(N*M)*(k) + i + N*(j-1)];} + if (k <= 0) {val3 = 0.0f;} else {val3 = R3[(N*M)*(k-1) + i + N*j];} + //Write final result to global memory + D[index] = Ad[index] - lambda*(R1[index] + R2[index] + R3[index] - val1 - val2 - val3); + } + return; +} + +__global__ void Grad_func3D_kernel(float *P1, float *P2, float *P3, float *D, float *R1, float *R2, float *R3, int N, int M, int Z, int ImSize, float multip) +{ + + float val1,val2,val3; + + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + /* boundary conditions */ + if (i >= N-1) val1 = 0.0f; else val1 = D[index] - D[(N*M)*(k) + (i+1) + N*j]; + if (j >= M-1) val2 = 0.0f; else val2 = D[index] - D[(N*M)*(k) + i + N*(j+1)]; + if (k >= Z-1) val3 = 0.0f; else val3 = D[index] - D[(N*M)*(k+1) + i + N*j]; + + //Write final result to global memory + P1[index] = R1[index] + multip*val1; + P2[index] = R2[index] + multip*val2; + P3[index] = R3[index] + multip*val3; + } + return; +} + +__global__ void Proj_func3D_iso_kernel(float *P1, float *P2, float *P3, int N, int M, int Z, int ImSize) +{ + + float denom,sq_denom; + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + denom = pow(P1[index],2) + pow(P2[index],2) + pow(P3[index],2); + + if (denom > 1.0f) { + sq_denom = 1.0f/sqrt(denom); + P1[index] = P1[index]*sq_denom; + P2[index] = P2[index]*sq_denom; + P3[index] = P3[index]*sq_denom; + } + } + return; +} + +__global__ void Proj_func3D_aniso_kernel(float *P1, float *P2, float *P3, int N, int M, int Z, int ImSize) +{ + + float val1, val2, val3; + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + val1 = abs(P1[index]); + val2 = abs(P2[index]); + val3 = abs(P3[index]); + if (val1 < 1.0f) {val1 = 1.0f;} + if (val2 < 1.0f) {val2 = 1.0f;} + if (val3 < 1.0f) {val3 = 1.0f;} + P1[index] = P1[index]/val1; + P2[index] = P2[index]/val2; + P3[index] = P3[index]/val3; + } + return; +} +__global__ void Rupd_func3D_kernel(float *P1, float *P1_old, float *P2, float *P2_old, float *P3, float *P3_old, float *R1, float *R2, float *R3, float tkp1, float tk, float multip2, int N, int M, int Z, int ImSize) +{ + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + R1[index] = P1[index] + multip2*(P1[index] - P1_old[index]); + R2[index] = P2[index] + multip2*(P2[index] - P2_old[index]); + R3[index] = P3[index] + multip2*(P3[index] - P3_old[index]); + } + return; +} + +__global__ void nonneg3D_kernel(float* Output, int N, int M, int Z, int num_total) +{ + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if (index < num_total) { + if (Output[index] < 0.0f) Output[index] = 0.0f; + } +} +__global__ void FGPcopy_kernel2D(float *Input, float* Output, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + Output[index] = Input[index]; + } +} + +__global__ void FGPcopy_kernel3D(float *Input, float* Output, int N, int M, int Z, int num_total) +{ + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if (index < num_total) { + Output[index] = Input[index]; + } +} + +__global__ void FGPResidCalc2D_kernel(float *Input1, float *Input2, float* Output, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + Output[index] = Input1[index] - Input2[index]; + } +} + +__global__ void FGPResidCalc3D_kernel(float *Input1, float *Input2, float* Output, int N, int M, int Z, int num_total) +{ + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if (index < num_total) { + Output[index] = Input1[index] - Input2[index]; + } +} + +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ + +////////////MAIN HOST FUNCTION /////////////// +extern "C" int TV_FGP_GPU_main(float *Input, float *Output, float lambdaPar, int iter, float epsil, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ) +{ + int deviceCount = -1; // number of devices + cudaGetDeviceCount(&deviceCount); + if (deviceCount == 0) { + fprintf(stderr, "No CUDA devices found\n"); + return -1; + } + + int count = 0, i; + float re, multip,multip2; + float tk = 1.0f; + float tkp1=1.0f; + + if (dimZ <= 1) { + /*2D verson*/ + int ImSize = dimX*dimY; + float *d_input, *d_update=NULL, *d_update_prev=NULL, *P1=NULL, *P2=NULL, *P1_prev=NULL, *P2_prev=NULL, *R1=NULL, *R2=NULL; + + dim3 dimBlock(BLKXSIZE2D,BLKYSIZE2D); + dim3 dimGrid(idivup(dimX,BLKXSIZE2D), idivup(dimY,BLKYSIZE2D)); + + /*allocate space for images on device*/ + checkCudaErrors( cudaMalloc((void**)&d_input,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_update,ImSize*sizeof(float)) ); + if (epsil != 0.0f) checkCudaErrors( cudaMalloc((void**)&d_update_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P1,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P2,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P1_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P2_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R1,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R2,ImSize*sizeof(float)) ); + + checkCudaErrors( cudaMemcpy(d_input,Input,ImSize*sizeof(float),cudaMemcpyHostToDevice)); + cudaMemset(P1, 0, ImSize*sizeof(float)); + cudaMemset(P2, 0, ImSize*sizeof(float)); + cudaMemset(P1_prev, 0, ImSize*sizeof(float)); + cudaMemset(P2_prev, 0, ImSize*sizeof(float)); + cudaMemset(R1, 0, ImSize*sizeof(float)); + cudaMemset(R2, 0, ImSize*sizeof(float)); + + /********************** Run CUDA 2D kernel here ********************/ + multip = (1.0f/(8.0f*lambdaPar)); + + /* The main kernel */ + for (i = 0; i < iter; i++) { + + /* computing the gradient of the objective function */ + Obj_func2D_kernel<<<dimGrid,dimBlock>>>(d_input, d_update, R1, R2, dimX, dimY, ImSize, lambdaPar); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + if (nonneg != 0) { + nonneg2D_kernel<<<dimGrid,dimBlock>>>(d_update, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); } + + /*Taking a step towards minus of the gradient*/ + Grad_func2D_kernel<<<dimGrid,dimBlock>>>(P1, P2, d_update, R1, R2, dimX, dimY, ImSize, multip); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* projection step */ + if (methodTV == 0) Proj_func2D_iso_kernel<<<dimGrid,dimBlock>>>(P1, P2, dimX, dimY, ImSize); /*isotropic TV*/ + else Proj_func2D_aniso_kernel<<<dimGrid,dimBlock>>>(P1, P2, dimX, dimY, ImSize); /*anisotropic TV*/ + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + tkp1 = (1.0f + sqrt(1.0f + 4.0f*tk*tk))*0.5f; + multip2 = ((tk-1.0f)/tkp1); + + Rupd_func2D_kernel<<<dimGrid,dimBlock>>>(P1, P1_prev, P2, P2_prev, R1, R2, tkp1, tk, multip2, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + if (epsil != 0.0f) { + /* calculate norm - stopping rules using the Thrust library */ + FGPResidCalc2D_kernel<<<dimGrid,dimBlock>>>(d_update, d_update_prev, P1_prev, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + thrust::device_vector<float> d_vec(P1_prev, P1_prev + ImSize); + float reduction = sqrt(thrust::transform_reduce(d_vec.begin(), d_vec.end(), square(), 0.0f, thrust::plus<float>())); + thrust::device_vector<float> d_vec2(d_update, d_update + ImSize); + float reduction2 = sqrt(thrust::transform_reduce(d_vec2.begin(), d_vec2.end(), square(), 0.0f, thrust::plus<float>())); + + re = (reduction/reduction2); + if (re < epsil) count++; + if (count > 4) break; + + FGPcopy_kernel2D<<<dimGrid,dimBlock>>>(d_update, d_update_prev, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + } + + FGPcopy_kernel2D<<<dimGrid,dimBlock>>>(P1, P1_prev, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + FGPcopy_kernel2D<<<dimGrid,dimBlock>>>(P2, P2_prev, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + tk = tkp1; + } + if (printM == 1) printf("FGP-TV iterations stopped at iteration %i \n", i); + /***************************************************************/ + //copy result matrix from device to host memory + cudaMemcpy(Output,d_update,ImSize*sizeof(float),cudaMemcpyDeviceToHost); + + cudaFree(d_input); + cudaFree(d_update); + if (epsil != 0.0f) cudaFree(d_update_prev); + cudaFree(P1); + cudaFree(P2); + cudaFree(P1_prev); + cudaFree(P2_prev); + cudaFree(R1); + cudaFree(R2); + } + else { + /*3D verson*/ + int ImSize = dimX*dimY*dimZ; + float *d_input, *d_update=NULL, *P1=NULL, *P2=NULL, *P3=NULL, *P1_prev=NULL, *P2_prev=NULL, *P3_prev=NULL, *R1=NULL, *R2=NULL, *R3=NULL; + + dim3 dimBlock(BLKXSIZE,BLKYSIZE,BLKZSIZE); + dim3 dimGrid(idivup(dimX,BLKXSIZE), idivup(dimY,BLKYSIZE),idivup(dimZ,BLKZSIZE)); + + /*allocate space for images on device*/ + checkCudaErrors( cudaMalloc((void**)&d_input,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_update,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P1,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P2,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P3,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P1_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P2_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P3_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R1,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R2,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R3,ImSize*sizeof(float)) ); + + checkCudaErrors( cudaMemcpy(d_input,Input,ImSize*sizeof(float),cudaMemcpyHostToDevice)); + cudaMemset(P1, 0, ImSize*sizeof(float)); + cudaMemset(P2, 0, ImSize*sizeof(float)); + cudaMemset(P3, 0, ImSize*sizeof(float)); + cudaMemset(P1_prev, 0, ImSize*sizeof(float)); + cudaMemset(P2_prev, 0, ImSize*sizeof(float)); + cudaMemset(P3_prev, 0, ImSize*sizeof(float)); + cudaMemset(R1, 0, ImSize*sizeof(float)); + cudaMemset(R2, 0, ImSize*sizeof(float)); + cudaMemset(R3, 0, ImSize*sizeof(float)); + /********************** Run CUDA 3D kernel here ********************/ + multip = (1.0f/(26.0f*lambdaPar)); + + /* The main kernel */ + for (i = 0; i < iter; i++) { + + /* computing the gradient of the objective function */ + Obj_func3D_kernel<<<dimGrid,dimBlock>>>(d_input, d_update, R1, R2, R3, dimX, dimY, dimZ, ImSize, lambdaPar); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + if (nonneg != 0) { + nonneg3D_kernel<<<dimGrid,dimBlock>>>(d_update, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); } + + /*Taking a step towards minus of the gradient*/ + Grad_func3D_kernel<<<dimGrid,dimBlock>>>(P1, P2, P3, d_update, R1, R2, R3, dimX, dimY, dimZ, ImSize, multip); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* projection step */ + if (methodTV == 0) Proj_func3D_iso_kernel<<<dimGrid,dimBlock>>>(P1, P2, P3, dimX, dimY, dimZ, ImSize); /* isotropic kernel */ + else Proj_func3D_aniso_kernel<<<dimGrid,dimBlock>>>(P1, P2, P3, dimX, dimY, dimZ, ImSize); /* anisotropic kernel */ + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + tkp1 = (1.0f + sqrt(1.0f + 4.0f*tk*tk))*0.5f; + multip2 = ((tk-1.0f)/tkp1); + + Rupd_func3D_kernel<<<dimGrid,dimBlock>>>(P1, P1_prev, P2, P2_prev, P3, P3_prev, R1, R2, R3, tkp1, tk, multip2, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + FGPcopy_kernel3D<<<dimGrid,dimBlock>>>(P1, P1_prev, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + FGPcopy_kernel3D<<<dimGrid,dimBlock>>>(P2, P2_prev, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + FGPcopy_kernel3D<<<dimGrid,dimBlock>>>(P3, P3_prev, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + tk = tkp1; + } + if (printM == 1) printf("FGP-TV iterations stopped at iteration %i \n", i); + /***************************************************************/ + //copy result matrix from device to host memory + cudaMemcpy(Output,d_update,ImSize*sizeof(float),cudaMemcpyDeviceToHost); + + cudaFree(d_input); + cudaFree(d_update); + cudaFree(P1); + cudaFree(P2); + cudaFree(P3); + cudaFree(P1_prev); + cudaFree(P2_prev); + cudaFree(P3_prev); + cudaFree(R1); + cudaFree(R2); + cudaFree(R3); + } + //cudaDeviceReset(); + return 0; +} diff --git a/src/Core/regularisers_GPU/TV_FGP_GPU_core.h b/src/Core/regularisers_GPU/TV_FGP_GPU_core.h new file mode 100755 index 0000000..bf13508 --- /dev/null +++ b/src/Core/regularisers_GPU/TV_FGP_GPU_core.h @@ -0,0 +1,9 @@ +#ifndef _TV_FGP_GPU_ +#define _TV_FGP_GPU_ + +#include "CCPiDefines.h" +#include <memory.h> + +extern "C" CCPI_EXPORT int TV_FGP_GPU_main(float *Input, float *Output, float lambdaPar, int iter, float epsil, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ); + +#endif diff --git a/src/Core/regularisers_GPU/TV_ROF_GPU_core.cu b/src/Core/regularisers_GPU/TV_ROF_GPU_core.cu new file mode 100755 index 0000000..76f5be9 --- /dev/null +++ b/src/Core/regularisers_GPU/TV_ROF_GPU_core.cu @@ -0,0 +1,358 @@ + /* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "TV_ROF_GPU_core.h" + +/* C-OMP implementation of ROF-TV denoising/regularization model [1] (2D/3D case) +* +* Input Parameters: +* 1. Noisy image/volume [REQUIRED] +* 2. lambda - regularization parameter [REQUIRED] +* 3. tau - marching step for explicit scheme, ~0.1 is recommended [REQUIRED] +* 4. Number of iterations, for explicit scheme >= 150 is recommended [REQUIRED] +* +* Output: +* [1] Regularized image/volume + + * This function is based on the paper by +* [1] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" +* +* D. Kazantsev, 2016-18 +*/ +#include "shared.h" + +#define BLKXSIZE 8 +#define BLKYSIZE 8 +#define BLKZSIZE 8 + +#define BLKXSIZE2D 16 +#define BLKYSIZE2D 16 +#define EPS 1.0e-12 + +#define idivup(a, b) ( ((a)%(b) != 0) ? (a)/(b)+1 : (a)/(b) ) + +#define MAX(x, y) (((x) > (y)) ? (x) : (y)) +#define MIN(x, y) (((x) < (y)) ? (x) : (y)) + +__host__ __device__ int sign (float x) +{ + return (x > 0) - (x < 0); +} + +/*********************2D case****************************/ + + /* differences 1 */ + __global__ void D1_func2D(float* Input, float* D1, int N, int M) + { + int i1, j1, i2; + float NOMx_1,NOMy_1,NOMy_0,denom1,denom2,T1; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + N*j; + + if ((i >= 0) && (i < N) && (j >= 0) && (j < M)) { + + /* boundary conditions (Neumann reflections) */ + i1 = i + 1; if (i1 >= N) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= M) j1 = j-1; + + /* Forward-backward differences */ + NOMx_1 = Input[j1*N + i] - Input[index]; /* x+ */ + NOMy_1 = Input[j*N + i1] - Input[index]; /* y+ */ + NOMy_0 = Input[index] - Input[j*N + i2]; /* y- */ + + denom1 = NOMx_1*NOMx_1; + denom2 = 0.5f*(sign((float)NOMy_1) + sign((float)NOMy_0))*(MIN(abs((float)NOMy_1), abs((float)NOMy_0))); + denom2 = denom2*denom2; + T1 = sqrt(denom1 + denom2 + EPS); + D1[index] = NOMx_1/T1; + } + } + + /* differences 2 */ + __global__ void D2_func2D(float* Input, float* D2, int N, int M) + { + int i1, j1, j2; + float NOMx_1,NOMy_1,NOMx_0,denom1,denom2,T2; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + N*j; + + if ((i >= 0) && (i < (N)) && (j >= 0) && (j < (M))) { + + /* boundary conditions (Neumann reflections) */ + i1 = i + 1; if (i1 >= N) i1 = i-1; + j1 = j + 1; if (j1 >= M) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + + /* Forward-backward differences */ + NOMx_1 = Input[j1*N + i] - Input[index]; /* x+ */ + NOMy_1 = Input[j*N + i1] - Input[index]; /* y+ */ + NOMx_0 = Input[index] - Input[j2*N + i]; /* x- */ + + denom1 = NOMy_1*NOMy_1; + denom2 = 0.5f*(sign((float)NOMx_1) + sign((float)NOMx_0))*(MIN(abs((float)NOMx_1), abs((float)NOMx_0))); + denom2 = denom2*denom2; + T2 = sqrt(denom1 + denom2 + EPS); + D2[index] = NOMy_1/T2; + } + } + + __global__ void TV_kernel2D(float *D1, float *D2, float *Update, float *Input, float lambda, float tau, int N, int M) + { + int i2, j2; + float dv1,dv2; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + + int index = i + N*j; + + if ((i >= 0) && (i < (N)) && (j >= 0) && (j < (M))) { + + /* boundary conditions (Neumann reflections) */ + i2 = i - 1; if (i2 < 0) i2 = i+1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + + /* divergence components */ + dv1 = D1[index] - D1[j2*N + i]; + dv2 = D2[index] - D2[j*N + i2]; + + Update[index] += tau*(2.0f*lambda*(dv1 + dv2) - (Update[index] - Input[index])); + + } + } +/*********************3D case****************************/ + + /* differences 1 */ + __global__ void D1_func3D(float* Input, float* D1, int dimX, int dimY, int dimZ) + { + float NOMx_1, NOMy_1, NOMy_0, NOMz_1, NOMz_0, denom1, denom2,denom3, T1; + int i1,i2,k1,j1,j2,k2; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (dimX*dimY)*k + j*dimX+i; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /* Forward-backward differences */ + NOMx_1 = Input[(dimX*dimY)*k + j1*dimX + i] - Input[index]; /* x+ */ + NOMy_1 = Input[(dimX*dimY)*k + j*dimX + i1] - Input[index]; /* y+ */ + NOMy_0 = Input[index] - Input[(dimX*dimY)*k + j*dimX + i2]; /* y- */ + + NOMz_1 = Input[(dimX*dimY)*k1 + j*dimX + i] - Input[index]; /* z+ */ + NOMz_0 = Input[index] - Input[(dimX*dimY)*k2 + j*dimX + i]; /* z- */ + + + denom1 = NOMx_1*NOMx_1; + denom2 = 0.5*(sign(NOMy_1) + sign(NOMy_0))*(MIN(abs(NOMy_1),abs(NOMy_0))); + denom2 = denom2*denom2; + denom3 = 0.5*(sign(NOMz_1) + sign(NOMz_0))*(MIN(abs(NOMz_1),abs(NOMz_0))); + denom3 = denom3*denom3; + T1 = sqrt(denom1 + denom2 + denom3 + EPS); + D1[index] = NOMx_1/T1; + } + } + + /* differences 2 */ + __global__ void D2_func3D(float* Input, float* D2, int dimX, int dimY, int dimZ) + { + float NOMx_1, NOMy_1, NOMx_0, NOMz_1, NOMz_0, denom1, denom2, denom3, T2; + int i1,i2,k1,j1,j2,k2; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (dimX*dimY)*k + j*dimX+i; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + + /* Forward-backward differences */ + NOMx_1 = Input[(dimX*dimY)*k + (j1)*dimX + i] - Input[index]; /* x+ */ + NOMy_1 = Input[(dimX*dimY)*k + (j)*dimX + i1] - Input[index]; /* y+ */ + NOMx_0 = Input[index] - Input[(dimX*dimY)*k + (j2)*dimX + i]; /* x- */ + NOMz_1 = Input[(dimX*dimY)*k1 + j*dimX + i] - Input[index]; /* z+ */ + NOMz_0 = Input[index] - Input[(dimX*dimY)*k2 + (j)*dimX + i]; /* z- */ + + + denom1 = NOMy_1*NOMy_1; + denom2 = 0.5*(sign(NOMx_1) + sign(NOMx_0))*(MIN(abs(NOMx_1),abs(NOMx_0))); + denom2 = denom2*denom2; + denom3 = 0.5*(sign(NOMz_1) + sign(NOMz_0))*(MIN(abs(NOMz_1),abs(NOMz_0))); + denom3 = denom3*denom3; + T2 = sqrt(denom1 + denom2 + denom3 + EPS); + D2[index] = NOMy_1/T2; + } + } + + /* differences 3 */ + __global__ void D3_func3D(float* Input, float* D3, int dimX, int dimY, int dimZ) + { + float NOMx_1, NOMy_1, NOMx_0, NOMy_0, NOMz_1, denom1, denom2, denom3, T3; + int i1,i2,k1,j1,j2,k2; + + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (dimX*dimY)*k + j*dimX+i; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /* Forward-backward differences */ + NOMx_1 = Input[(dimX*dimY)*k + (j1)*dimX + i] - Input[index]; /* x+ */ + NOMy_1 = Input[(dimX*dimY)*k + (j)*dimX + i1] - Input[index]; /* y+ */ + NOMy_0 = Input[index] - Input[(dimX*dimY)*k + (j)*dimX + i2]; /* y- */ + NOMx_0 = Input[index] - Input[(dimX*dimY)*k + (j2)*dimX + i]; /* x- */ + NOMz_1 = Input[(dimX*dimY)*k1 + j*dimX + i] - Input[index]; /* z+ */ + + denom1 = NOMz_1*NOMz_1; + denom2 = 0.5*(sign(NOMx_1) + sign(NOMx_0))*(MIN(abs(NOMx_1),abs(NOMx_0))); + denom2 = denom2*denom2; + denom3 = 0.5*(sign(NOMy_1) + sign(NOMy_0))*(MIN(abs(NOMy_1),abs(NOMy_0))); + denom3 = denom3*denom3; + T3 = sqrt(denom1 + denom2 + denom3 + EPS); + D3[index] = NOMz_1/T3; + } + } + + __global__ void TV_kernel3D(float *D1, float *D2, float *D3, float *Update, float *Input, float lambda, float tau, int dimX, int dimY, int dimZ) + { + float dv1, dv2, dv3; + int i1,i2,k1,j1,j2,k2; + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (dimX*dimY)*k + j*dimX+i; + + if ((i >= 0) && (i < dimX) && (j >= 0) && (j < dimY) && (k >= 0) && (k < dimZ)) { + + /* symmetric boundary conditions (Neuman) */ + i1 = i + 1; if (i1 >= dimX) i1 = i-1; + i2 = i - 1; if (i2 < 0) i2 = i+1; + j1 = j + 1; if (j1 >= dimY) j1 = j-1; + j2 = j - 1; if (j2 < 0) j2 = j+1; + k1 = k + 1; if (k1 >= dimZ) k1 = k-1; + k2 = k - 1; if (k2 < 0) k2 = k+1; + + /*divergence components */ + dv1 = D1[index] - D1[(dimX*dimY)*k + j2*dimX+i]; + dv2 = D2[index] - D2[(dimX*dimY)*k + j*dimX+i2]; + dv3 = D3[index] - D3[(dimX*dimY)*k2 + j*dimX+i]; + + Update[index] += tau*(2.0f*lambda*(dv1 + dv2 + dv3) - (Update[index] - Input[index])); + + } + } + +///////////////////////////////////////////////// +// HOST FUNCTION +extern "C" int TV_ROF_GPU_main(float* Input, float* Output, float lambdaPar, int iter, float tau, int N, int M, int Z) +{ + // set up device + int dev = 0; + CHECK(cudaSetDevice(dev)); + float *d_input, *d_update, *d_D1, *d_D2; + + if (Z == 0) Z = 1; + CHECK(cudaMalloc((void**)&d_input,N*M*Z*sizeof(float))); + CHECK(cudaMalloc((void**)&d_update,N*M*Z*sizeof(float))); + CHECK(cudaMalloc((void**)&d_D1,N*M*Z*sizeof(float))); + CHECK(cudaMalloc((void**)&d_D2,N*M*Z*sizeof(float))); + + CHECK(cudaMemcpy(d_input,Input,N*M*Z*sizeof(float),cudaMemcpyHostToDevice)); + CHECK(cudaMemcpy(d_update,Input,N*M*Z*sizeof(float),cudaMemcpyHostToDevice)); + + if (Z > 1) { + // TV - 3D case + dim3 dimBlock(BLKXSIZE,BLKYSIZE,BLKZSIZE); + dim3 dimGrid(idivup(N,BLKXSIZE), idivup(M,BLKYSIZE),idivup(Z,BLKXSIZE)); + + float *d_D3; + CHECK(cudaMalloc((void**)&d_D3,N*M*Z*sizeof(float))); + + for(int n=0; n < iter; n++) { + /* calculate differences */ + D1_func3D<<<dimGrid,dimBlock>>>(d_update, d_D1, N, M, Z); + CHECK(cudaDeviceSynchronize()); + D2_func3D<<<dimGrid,dimBlock>>>(d_update, d_D2, N, M, Z); + CHECK(cudaDeviceSynchronize()); + D3_func3D<<<dimGrid,dimBlock>>>(d_update, d_D3, N, M, Z); + CHECK(cudaDeviceSynchronize()); + /*running main kernel*/ + TV_kernel3D<<<dimGrid,dimBlock>>>(d_D1, d_D2, d_D3, d_update, d_input, lambdaPar, tau, N, M, Z); + CHECK(cudaDeviceSynchronize()); + } + + CHECK(cudaFree(d_D3)); + } + else { + // TV - 2D case + dim3 dimBlock(BLKXSIZE2D,BLKYSIZE2D); + dim3 dimGrid(idivup(N,BLKXSIZE2D), idivup(M,BLKYSIZE2D)); + + for(int n=0; n < iter; n++) { + /* calculate differences */ + D1_func2D<<<dimGrid,dimBlock>>>(d_update, d_D1, N, M); + CHECK(cudaDeviceSynchronize()); + D2_func2D<<<dimGrid,dimBlock>>>(d_update, d_D2, N, M); + CHECK(cudaDeviceSynchronize()); + /*running main kernel*/ + TV_kernel2D<<<dimGrid,dimBlock>>>(d_D1, d_D2, d_update, d_input, lambdaPar, tau, N, M); + CHECK(cudaDeviceSynchronize()); + } + } + CHECK(cudaMemcpy(Output,d_update,N*M*Z*sizeof(float),cudaMemcpyDeviceToHost)); + CHECK(cudaFree(d_input)); + CHECK(cudaFree(d_update)); + CHECK(cudaFree(d_D1)); + CHECK(cudaFree(d_D2)); + //cudaDeviceReset(); + return 0; +} diff --git a/src/Core/regularisers_GPU/TV_ROF_GPU_core.h b/src/Core/regularisers_GPU/TV_ROF_GPU_core.h new file mode 100755 index 0000000..3a09296 --- /dev/null +++ b/src/Core/regularisers_GPU/TV_ROF_GPU_core.h @@ -0,0 +1,8 @@ +#ifndef __TVGPU_H__ +#define __TVGPU_H__ +#include "CCPiDefines.h" +#include <stdio.h> + +extern "C" CCPI_EXPORT int TV_ROF_GPU_main(float* Input, float* Output, float lambdaPar, int iter, float tau, int N, int M, int Z); + +#endif diff --git a/src/Core/regularisers_GPU/TV_SB_GPU_core.cu b/src/Core/regularisers_GPU/TV_SB_GPU_core.cu new file mode 100755 index 0000000..1f494ee --- /dev/null +++ b/src/Core/regularisers_GPU/TV_SB_GPU_core.cu @@ -0,0 +1,552 @@ + /* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "TV_SB_GPU_core.h" +#include "shared.h" +#include <thrust/device_vector.h> +#include <thrust/transform_reduce.h> + +/* CUDA implementation of Split Bregman - TV denoising-regularisation model (2D/3D) [1] +* +* Input Parameters: +* 1. Noisy image/volume +* 2. lambda - regularisation parameter +* 3. Number of iterations [OPTIONAL parameter] +* 4. eplsilon - tolerance constant [OPTIONAL parameter] +* 5. TV-type: 'iso' or 'l1' [OPTIONAL parameter] +* 6. nonneg: 'nonnegativity (0 is OFF by default) [OPTIONAL parameter] +* 7. print information: 0 (off) or 1 (on) [OPTIONAL parameter] +* +* Output: +* 1. Filtered/regularized image +* +* [1]. Goldstein, T. and Osher, S., 2009. The split Bregman method for L1-regularized problems. SIAM journal on imaging sciences, 2(2), pp.323-343. +*/ + +// This will output the proper CUDA error strings in the event that a CUDA host call returns an error + +#define BLKXSIZE2D 16 +#define BLKYSIZE2D 16 + +#define BLKXSIZE 8 +#define BLKYSIZE 8 +#define BLKZSIZE 8 + +#define idivup(a, b) ( ((a)%(b) != 0) ? (a)/(b)+1 : (a)/(b) ) +struct square { __host__ __device__ float operator()(float x) { return x * x; } }; + +/************************************************/ +/*****************2D modules*********************/ +/************************************************/ +__global__ void gauss_seidel2D_kernel(float *U, float *A, float *U_prev, float *Dx, float *Dy, float *Bx, float *By, float lambda, float mu, float normConst, int N, int M, int ImSize) +{ + + float sum; + int i1,i2,j1,j2; + + //calculate each thread global index + const int i=blockIdx.x*blockDim.x+threadIdx.x; + const int j=blockIdx.y*blockDim.y+threadIdx.y; + + int index = j*N+i; + + if ((i < N) && (j < M)) { + i1 = i+1; if (i1 == N) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == M) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + + sum = Dx[j*N+i2] - Dx[index] + Dy[j2*N+i] - Dy[index] - Bx[j*N+i2] + Bx[index] - By[j2*N+i] + By[index]; + sum += U_prev[j*N+i1] + U_prev[j*N+i2] + U_prev[j1*N+i] + U_prev[j2*N+i]; + sum *= lambda; + sum += mu*A[index]; + U[index] = normConst*sum; //Write final result to global memory + } + return; +} +__global__ void updDxDy_shrinkAniso2D_kernel(float *U, float *Dx, float *Dy, float *Bx, float *By, float lambda, int N, int M, int ImSize) +{ + + int i1,j1; + float val1, val11, val2, val22, denom_lam; + denom_lam = 1.0f/lambda; + + //calculate each thread global index + const int i=blockIdx.x*blockDim.x+threadIdx.x; + const int j=blockIdx.y*blockDim.y+threadIdx.y; + + int index = j*N+i; + + if ((i < N) && (j < M)) { + i1 = i+1; if (i1 == N) i1 = i-1; + j1 = j+1; if (j1 == M) j1 = j-1; + + val1 = (U[j*N+i1] - U[index]) + Bx[index]; + val2 = (U[j1*N+i] - U[index]) + By[index]; + + val11 = abs(val1) - denom_lam; if (val11 < 0) val11 = 0; + val22 = abs(val2) - denom_lam; if (val22 < 0) val22 = 0; + + if (val1 !=0) Dx[index] = (val1/abs(val1))*val11; else Dx[index] = 0; + if (val2 !=0) Dy[index] = (val2/abs(val2))*val22; else Dy[index] = 0; + } + return; +} + +__global__ void updDxDy_shrinkIso2D_kernel(float *U, float *Dx, float *Dy, float *Bx, float *By, float lambda, int N, int M, int ImSize) +{ + + int i1,j1; + float val1, val11, val2, denom_lam, denom; + denom_lam = 1.0f/lambda; + + //calculate each thread global index + const int i=blockIdx.x*blockDim.x+threadIdx.x; + const int j=blockIdx.y*blockDim.y+threadIdx.y; + + int index = j*N+i; + + if ((i < N) && (j < M)) { + i1 = i+1; if (i1 == N) i1 = i-1; + j1 = j+1; if (j1 == M) j1 = j-1; + + val1 = (U[j*N+i1] - U[index]) + Bx[index]; + val2 = (U[j1*N+i] - U[index]) + By[index]; + + denom = sqrt(val1*val1 + val2*val2); + + val11 = (denom - denom_lam); if (val11 < 0) val11 = 0.0f; + + if (denom != 0.0f) { + Dx[index] = val11*(val1/denom); + Dy[index] = val11*(val2/denom); + } + else { + Dx[index] = 0; + Dy[index] = 0; + } + } + return; +} + +__global__ void updBxBy2D_kernel(float *U, float *Dx, float *Dy, float *Bx, float *By, int N, int M, int ImSize) +{ + int i1,j1; + + //calculate each thread global index + const int i=blockIdx.x*blockDim.x+threadIdx.x; + const int j=blockIdx.y*blockDim.y+threadIdx.y; + + int index = j*N+i; + + if ((i < N) && (j < M)) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == N) i1 = i-1; + j1 = j+1; if (j1 == M) j1 = j-1; + + Bx[index] += (U[j*N+i1] - U[index]) - Dx[index]; + By[index] += (U[j1*N+i] - U[index]) - Dy[index]; + } + return; +} + + +/************************************************/ +/*****************3D modules*********************/ +/************************************************/ +__global__ void gauss_seidel3D_kernel(float *U, float *A, float *U_prev, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, float lambda, float mu, float normConst, int N, int M, int Z, int ImSize) +{ + + float sum,d_val,b_val; + int i1,i2,j1,j2,k1,k2; + + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + i1 = i+1; if (i1 == N) i1 = i-1; + i2 = i-1; if (i2 < 0) i2 = i+1; + j1 = j+1; if (j1 == M) j1 = j-1; + j2 = j-1; if (j2 < 0) j2 = j+1; + k1 = k+1; if (k1 == Z) k1 = k-1; + k2 = k-1; if (k2 < 0) k2 = k+1; + + d_val = Dx[(N*M)*k + j*N+i2] - Dx[index] + Dy[(N*M)*k + j2*N+i] - Dy[index] + Dz[(N*M)*k2 + j*N+i] - Dz[index]; + b_val = -Bx[(N*M)*k + j*N+i2] + Bx[index] - By[(N*M)*k + j2*N+i] + By[index] - Bz[(N*M)*k2 + j*N+i] + Bz[index]; + sum = d_val + b_val; + sum += U_prev[(N*M)*k + j*N+i1] + U_prev[(N*M)*k + j*N+i2] + U_prev[(N*M)*k + j1*N+i] + U_prev[(N*M)*k + j2*N+i] + U_prev[(N*M)*k1 + j*N+i] + U_prev[(N*M)*k2 + j*N+i]; + sum *= lambda; + sum += mu*A[index]; + U[index] = normConst*sum; + } + return; +} +__global__ void updDxDy_shrinkAniso3D_kernel(float *U, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, float lambda, int N, int M, int Z, int ImSize) +{ + + int i1,j1,k1; + float val1, val11, val2, val3, val22, val33, denom_lam; + denom_lam = 1.0f/lambda; + + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + i1 = i+1; if (i1 == N) i1 = i-1; + j1 = j+1; if (j1 == M) j1 = j-1; + k1 = k+1; if (k1 == Z) k1 = k-1; + + val1 = (U[(N*M)*k + i1 + N*j] - U[index]) + Bx[index]; + val2 = (U[(N*M)*k + i + N*j1] - U[index]) + By[index]; + val3 = (U[(N*M)*k1 + i + N*j] - U[index]) + Bz[index]; + + val11 = abs(val1) - denom_lam; if (val11 < 0.0f) val11 = 0.0f; + val22 = abs(val2) - denom_lam; if (val22 < 0.0f) val22 = 0.0f; + val33 = abs(val3) - denom_lam; if (val33 < 0.0f) val33 = 0.0f; + + if (val1 !=0.0f) Dx[index] = (val1/abs(val1))*val11; else Dx[index] = 0.0f; + if (val2 !=0.0f) Dy[index] = (val2/abs(val2))*val22; else Dy[index] = 0.0f; + if (val3 !=0.0f) Dz[index] = (val3/abs(val3))*val33; else Dz[index] = 0.0f; + } + return; +} + +__global__ void updDxDy_shrinkIso3D_kernel(float *U, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, float lambda, int N, int M, int Z, int ImSize) +{ + + int i1,j1,k1; + float val1, val11, val2, val3, denom_lam, denom; + denom_lam = 1.0f/lambda; + + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + i1 = i+1; if (i1 == N) i1 = i-1; + j1 = j+1; if (j1 == M) j1 = j-1; + k1 = k+1; if (k1 == Z) k1 = k-1; + + val1 = (U[(N*M)*k + i1 + N*j] - U[index]) + Bx[index]; + val2 = (U[(N*M)*k + i + N*j1] - U[index]) + By[index]; + val3 = (U[(N*M)*k1 + i + N*j] - U[index]) + Bz[index]; + + denom = sqrt(val1*val1 + val2*val2 + val3*val3); + + val11 = (denom - denom_lam); if (val11 < 0.0f) val11 = 0.0f; + + if (denom != 0.0f) { + Dx[index] = val11*(val1/denom); + Dy[index] = val11*(val2/denom); + Dz[index] = val11*(val3/denom); + } + else { + Dx[index] = 0.0f; + Dy[index] = 0.0f; + Dz[index] = 0.0f; + } + } + return; +} + +__global__ void updBxBy3D_kernel(float *U, float *Dx, float *Dy, float *Dz, float *Bx, float *By, float *Bz, int N, int M, int Z, int ImSize) +{ + int i1,j1,k1; + + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + /* symmetric boundary conditions (Neuman) */ + i1 = i+1; if (i1 == N) i1 = i-1; + j1 = j+1; if (j1 == M) j1 = j-1; + k1 = k+1; if (k1 == Z) k1 = k-1; + + Bx[index] += (U[(N*M)*k + i1 + N*j] - U[index]) - Dx[index]; + By[index] += (U[(N*M)*k + i + N*j1] - U[index]) - Dy[index]; + Bz[index] += (U[(N*M)*k1 + i + N*j] - U[index]) - Dz[index]; + } + return; +} + +__global__ void SBcopy_kernel2D(float *Input, float* Output, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + Output[index] = Input[index]; + } +} + +__global__ void SBcopy_kernel3D(float *Input, float* Output, int N, int M, int Z, int num_total) +{ + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if (index < num_total) { + Output[index] = Input[index]; + } +} + +__global__ void SBResidCalc2D_kernel(float *Input1, float *Input2, float* Output, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + Output[index] = Input1[index] - Input2[index]; + } +} + +__global__ void SBResidCalc3D_kernel(float *Input1, float *Input2, float* Output, int N, int M, int Z, int num_total) +{ + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if (index < num_total) { + Output[index] = Input1[index] - Input2[index]; + } +} + +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ +/********************* MAIN HOST FUNCTION ******************/ +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ +extern "C" int TV_SB_GPU_main(float *Input, float *Output, float mu, int iter, float epsil, int methodTV, int printM, int dimX, int dimY, int dimZ) +{ + int deviceCount = -1; // number of devices + cudaGetDeviceCount(&deviceCount); + if (deviceCount == 0) { + fprintf(stderr, "No CUDA devices found\n"); + return -1; + } + + int ll, DimTotal; + float re, lambda, normConst; + int count = 0; + mu = 1.0f/mu; + lambda = 2.0f*mu; + + if (dimZ <= 1) { + /*2D verson*/ + DimTotal = dimX*dimY; + normConst = 1.0f/(mu + 4.0f*lambda); + float *d_input, *d_update, *d_res, *d_update_prev=NULL, *Dx=NULL, *Dy=NULL, *Bx=NULL, *By=NULL; + + dim3 dimBlock(BLKXSIZE2D,BLKYSIZE2D); + dim3 dimGrid(idivup(dimX,BLKXSIZE2D), idivup(dimY,BLKYSIZE2D)); + + /*allocate space for images on device*/ + checkCudaErrors( cudaMalloc((void**)&d_input,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_update,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_update_prev,DimTotal*sizeof(float)) ); + if (epsil != 0.0f) checkCudaErrors( cudaMalloc((void**)&d_res,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&Dx,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&Dy,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&Bx,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&By,DimTotal*sizeof(float)) ); + + checkCudaErrors( cudaMemcpy(d_input,Input,DimTotal*sizeof(float),cudaMemcpyHostToDevice)); + checkCudaErrors( cudaMemcpy(d_update,Input,DimTotal*sizeof(float),cudaMemcpyHostToDevice)); + cudaMemset(Dx, 0, DimTotal*sizeof(float)); + cudaMemset(Dy, 0, DimTotal*sizeof(float)); + cudaMemset(Bx, 0, DimTotal*sizeof(float)); + cudaMemset(By, 0, DimTotal*sizeof(float)); + + /********************** Run CUDA 2D kernels here ********************/ + /* The main kernel */ + for (ll = 0; ll < iter; ll++) { + + /* storing old value */ + SBcopy_kernel2D<<<dimGrid,dimBlock>>>(d_update, d_update_prev, dimX, dimY, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* perform two GS iterations (normally 2 is enough for the convergence) */ + gauss_seidel2D_kernel<<<dimGrid,dimBlock>>>(d_update, d_input, d_update_prev, Dx, Dy, Bx, By, lambda, mu, normConst, dimX, dimY, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + SBcopy_kernel2D<<<dimGrid,dimBlock>>>(d_update, d_update_prev, dimX, dimY, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + /* 2nd GS iteration */ + gauss_seidel2D_kernel<<<dimGrid,dimBlock>>>(d_update, d_input, d_update_prev, Dx, Dy, Bx, By, lambda, mu, normConst, dimX, dimY, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* TV-related step */ + if (methodTV == 1) updDxDy_shrinkAniso2D_kernel<<<dimGrid,dimBlock>>>(d_update, Dx, Dy, Bx, By, lambda, dimX, dimY, DimTotal); + else updDxDy_shrinkIso2D_kernel<<<dimGrid,dimBlock>>>(d_update, Dx, Dy, Bx, By, lambda, dimX, dimY, DimTotal); + + /* update for Bregman variables */ + updBxBy2D_kernel<<<dimGrid,dimBlock>>>(d_update, Dx, Dy, Bx, By, dimX, dimY, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + if (epsil != 0.0f) { + /* calculate norm - stopping rules using the Thrust library */ + SBResidCalc2D_kernel<<<dimGrid,dimBlock>>>(d_update, d_update_prev, d_res, dimX, dimY, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + thrust::device_vector<float> d_vec(d_res, d_res + DimTotal); + float reduction = sqrt(thrust::transform_reduce(d_vec.begin(), d_vec.end(), square(), 0.0f, thrust::plus<float>())); + thrust::device_vector<float> d_vec2(d_update, d_update + DimTotal); + float reduction2 = sqrt(thrust::transform_reduce(d_vec2.begin(), d_vec2.end(), square(), 0.0f, thrust::plus<float>())); + + re = (reduction/reduction2); + if (re < epsil) count++; + if (count > 4) break; + } + + } + if (printM == 1) printf("SB-TV iterations stopped at iteration %i \n", ll); + /***************************************************************/ + //copy result matrix from device to host memory + cudaMemcpy(Output,d_update,DimTotal*sizeof(float),cudaMemcpyDeviceToHost); + + cudaFree(d_input); + cudaFree(d_update); + cudaFree(d_update_prev); + if (epsil != 0.0f) cudaFree(d_res); + cudaFree(Dx); + cudaFree(Dy); + cudaFree(Bx); + cudaFree(By); + } + else { + /*3D verson*/ + DimTotal = dimX*dimY*dimZ; + normConst = 1.0f/(mu + 6.0f*lambda); + float *d_input, *d_update, *d_res, *d_update_prev=NULL, *Dx=NULL, *Dy=NULL, *Dz=NULL, *Bx=NULL, *By=NULL, *Bz=NULL; + + dim3 dimBlock(BLKXSIZE,BLKYSIZE,BLKZSIZE); + dim3 dimGrid(idivup(dimX,BLKXSIZE), idivup(dimY,BLKYSIZE),idivup(dimZ,BLKZSIZE)); + + /*allocate space for images on device*/ + checkCudaErrors( cudaMalloc((void**)&d_input,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_update,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_update_prev,DimTotal*sizeof(float)) ); + if (epsil != 0.0f) checkCudaErrors( cudaMalloc((void**)&d_res,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&Dx,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&Dy,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&Dz,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&Bx,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&By,DimTotal*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&Bz,DimTotal*sizeof(float)) ); + + checkCudaErrors( cudaMemcpy(d_input,Input,DimTotal*sizeof(float),cudaMemcpyHostToDevice)); + checkCudaErrors( cudaMemcpy(d_update,Input,DimTotal*sizeof(float),cudaMemcpyHostToDevice)); + cudaMemset(Dx, 0, DimTotal*sizeof(float)); + cudaMemset(Dy, 0, DimTotal*sizeof(float)); + cudaMemset(Dz, 0, DimTotal*sizeof(float)); + cudaMemset(Bx, 0, DimTotal*sizeof(float)); + cudaMemset(By, 0, DimTotal*sizeof(float)); + cudaMemset(Bz, 0, DimTotal*sizeof(float)); + + /********************** Run CUDA 3D kernels here ********************/ + /* The main kernel */ + for (ll = 0; ll < iter; ll++) { + + /* storing old value */ + SBcopy_kernel3D<<<dimGrid,dimBlock>>>(d_update, d_update_prev, dimX, dimY, dimZ, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* perform two GS iterations (normally 2 is enough for the convergence) */ + gauss_seidel3D_kernel<<<dimGrid,dimBlock>>>(d_update, d_input, d_update_prev, Dx, Dy, Dz, Bx, By, Bz, lambda, mu, normConst, dimX, dimY, dimZ, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + SBcopy_kernel3D<<<dimGrid,dimBlock>>>(d_update, d_update_prev, dimX, dimY, dimZ, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + /* 2nd GS iteration */ + gauss_seidel3D_kernel<<<dimGrid,dimBlock>>>(d_update, d_input, d_update_prev, Dx, Dy, Dz, Bx, By, Bz, lambda, mu, normConst, dimX, dimY, dimZ, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* TV-related step */ + if (methodTV == 1) updDxDy_shrinkAniso3D_kernel<<<dimGrid,dimBlock>>>(d_update, Dx, Dy, Dz, Bx, By, Bz, lambda, dimX, dimY, dimZ, DimTotal); + else updDxDy_shrinkIso3D_kernel<<<dimGrid,dimBlock>>>(d_update, Dx, Dy, Dz, Bx, By, Bz, lambda, dimX, dimY, dimZ, DimTotal); + + /* update for Bregman variables */ + updBxBy3D_kernel<<<dimGrid,dimBlock>>>(d_update, Dx, Dy, Dz, Bx, By, Bz, dimX, dimY, dimZ, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + if (epsil != 0.0f) { + /* calculate norm - stopping rules using the Thrust library */ + SBResidCalc3D_kernel<<<dimGrid,dimBlock>>>(d_update, d_update_prev, d_res, dimX, dimY, dimZ, DimTotal); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + thrust::device_vector<float> d_vec(d_res, d_res + DimTotal); + float reduction = sqrt(thrust::transform_reduce(d_vec.begin(), d_vec.end(), square(), 0.0f, thrust::plus<float>())); + thrust::device_vector<float> d_vec2(d_update, d_update + DimTotal); + float reduction2 = sqrt(thrust::transform_reduce(d_vec2.begin(), d_vec2.end(), square(), 0.0f, thrust::plus<float>())); + + re = (reduction/reduction2); + if (re < epsil) count++; + if (count > 4) break; + } + } + if (printM == 1) printf("SB-TV iterations stopped at iteration %i \n", ll); + /***************************************************************/ + //copy result matrix from device to host memory + cudaMemcpy(Output,d_update,DimTotal*sizeof(float),cudaMemcpyDeviceToHost); + + cudaFree(d_input); + cudaFree(d_update); + cudaFree(d_update_prev); + if (epsil != 0.0f) cudaFree(d_res); + cudaFree(Dx); + cudaFree(Dy); + cudaFree(Dz); + cudaFree(Bx); + cudaFree(By); + cudaFree(Bz); + } + //cudaDeviceReset(); + return 0; +} diff --git a/src/Core/regularisers_GPU/TV_SB_GPU_core.h b/src/Core/regularisers_GPU/TV_SB_GPU_core.h new file mode 100755 index 0000000..901b90f --- /dev/null +++ b/src/Core/regularisers_GPU/TV_SB_GPU_core.h @@ -0,0 +1,10 @@ +#ifndef _SB_TV_GPU_ +#define _SB_TV_GPU_ + +#include "CCPiDefines.h" +#include <memory.h> + + +extern "C" CCPI_EXPORT int TV_SB_GPU_main(float *Input, float *Output, float mu, int iter, float epsil, int methodTV, int printM, int dimX, int dimY, int dimZ); + +#endif diff --git a/src/Core/regularisers_GPU/dTV_FGP_GPU_core.cu b/src/Core/regularisers_GPU/dTV_FGP_GPU_core.cu new file mode 100644 index 0000000..7503ec7 --- /dev/null +++ b/src/Core/regularisers_GPU/dTV_FGP_GPU_core.cu @@ -0,0 +1,741 @@ + /* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ +#include "shared.h" +#include "dTV_FGP_GPU_core.h" +#include <thrust/device_vector.h> +#include <thrust/transform_reduce.h> + +/* CUDA implementation of FGP-dTV [1,2] denoising/regularization model (2D/3D case) + * which employs structural similarity of the level sets of two images/volumes, see [1,2] + * The current implementation updates image 1 while image 2 is being fixed. + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. Additional reference image/volume of the same dimensions as (1) [REQUIRED] + * 3. lambdaPar - regularization parameter [REQUIRED] + * 4. Number of iterations [OPTIONAL] + * 5. eplsilon: tolerance constant [OPTIONAL] + * 6. eta: smoothing constant to calculate gradient of the reference [OPTIONAL] * + * 7. TV-type: methodTV - 'iso' (0) or 'l1' (1) [OPTIONAL] + * 8. nonneg: 'nonnegativity (0 is OFF by default) [OPTIONAL] + * 9. print information: 0 (off) or 1 (on) [OPTIONAL] + * + * Output: + * [1] Filtered/regularized image/volume + * + * This function is based on the Matlab's codes and papers by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + * [2] M. J. Ehrhardt and M. M. Betcke, Multi-Contrast MRI Reconstruction with Structure-Guided Total Variation, SIAM Journal on Imaging Sciences 9(3), pp. 1084–1106 + */ + + +#define BLKXSIZE2D 16 +#define BLKYSIZE2D 16 + +#define BLKXSIZE 8 +#define BLKYSIZE 8 +#define BLKZSIZE 8 + +#define idivup(a, b) ( ((a)%(b) != 0) ? (a)/(b)+1 : (a)/(b) ) +struct square { __host__ __device__ float operator()(float x) { return x * x; } }; + +/************************************************/ +/*****************2D modules*********************/ +/************************************************/ + +__global__ void GradNorm_func2D_kernel(float *Refd, float *Refd_x, float *Refd_y, float eta, int N, int M, int ImSize) +{ + + float val1, val2, gradX, gradY, magn; + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + /* boundary conditions */ + if (xIndex >= N-1) val1 = 0.0f; else val1 = Refd[(xIndex+1) + N*yIndex]; + if (yIndex >= M-1) val2 = 0.0f; else val2 = Refd[(xIndex) + N*(yIndex + 1)]; + + gradX = val1 - Refd[index]; + gradY = val2 - Refd[index]; + magn = pow(gradX,2) + pow(gradY,2); + magn = sqrt(magn + pow(eta,2)); + Refd_x[index] = gradX/magn; + Refd_y[index] = gradY/magn; + } + return; +} + +__global__ void ProjectVect_func2D_kernel(float *R1, float *R2, float *Refd_x, float *Refd_y, int N, int M, int ImSize) +{ + + float in_prod; + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + in_prod = R1[index]*Refd_x[index] + R2[index]*Refd_y[index]; /* calculate inner product */ + R1[index] = R1[index] - in_prod*Refd_x[index]; + R2[index] = R2[index] - in_prod*Refd_y[index]; + } + return; +} + + +__global__ void Obj_dfunc2D_kernel(float *Ad, float *D, float *R1, float *R2, int N, int M, int ImSize, float lambda) +{ + + float val1,val2; + + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + if (xIndex <= 0) {val1 = 0.0f;} else {val1 = R1[(xIndex-1) + N*yIndex];} + if (yIndex <= 0) {val2 = 0.0f;} else {val2 = R2[xIndex + N*(yIndex-1)];} + + //Write final result to global memory + D[index] = Ad[index] - lambda*(R1[index] + R2[index] - val1 - val2); + } + return; +} + +__global__ void Grad_dfunc2D_kernel(float *P1, float *P2, float *D, float *R1, float *R2, float *Refd_x, float *Refd_y, int N, int M, int ImSize, float multip) +{ + + float val1,val2,in_prod; + + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + + /* boundary conditions */ + if (xIndex >= N-1) val1 = 0.0f; else val1 = D[index] - D[(xIndex+1) + N*yIndex]; + if (yIndex >= M-1) val2 = 0.0f; else val2 = D[index] - D[(xIndex) + N*(yIndex + 1)]; + + in_prod = val1*Refd_x[index] + val2*Refd_y[index]; /* calculate inner product */ + val1 = val1 - in_prod*Refd_x[index]; + val2 = val2 - in_prod*Refd_y[index]; + + //Write final result to global memory + P1[index] = R1[index] + multip*val1; + P2[index] = R2[index] + multip*val2; + } + return; +} + +__global__ void Proj_dfunc2D_iso_kernel(float *P1, float *P2, int N, int M, int ImSize) +{ + + float denom; + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + denom = pow(P1[index],2) + pow(P2[index],2); + if (denom > 1.0f) { + P1[index] = P1[index]/sqrt(denom); + P2[index] = P2[index]/sqrt(denom); + } + } + return; +} +__global__ void Proj_dfunc2D_aniso_kernel(float *P1, float *P2, int N, int M, int ImSize) +{ + + float val1, val2; + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + val1 = abs(P1[index]); + val2 = abs(P2[index]); + if (val1 < 1.0f) {val1 = 1.0f;} + if (val2 < 1.0f) {val2 = 1.0f;} + P1[index] = P1[index]/val1; + P2[index] = P2[index]/val2; + } + return; +} +__global__ void Rupd_dfunc2D_kernel(float *P1, float *P1_old, float *P2, float *P2_old, float *R1, float *R2, float tkp1, float tk, float multip2, int N, int M, int ImSize) +{ + //calculate each thread global index + const int xIndex=blockIdx.x*blockDim.x+threadIdx.x; + const int yIndex=blockIdx.y*blockDim.y+threadIdx.y; + + int index = xIndex + N*yIndex; + + if ((xIndex < N) && (yIndex < M)) { + R1[index] = P1[index] + multip2*(P1[index] - P1_old[index]); + R2[index] = P2[index] + multip2*(P2[index] - P2_old[index]); + } + return; +} +__global__ void dTVnonneg2D_kernel(float* Output, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + if (Output[index] < 0.0f) Output[index] = 0.0f; + } +} +__global__ void dTVcopy_kernel2D(float *Input, float* Output, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + Output[index] = Input[index]; + } +} + +__global__ void dTVcopy_kernel3D(float *Input, float* Output, int N, int M, int Z, int num_total) +{ + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if (index < num_total) { + Output[index] = Input[index]; + } +} + +__global__ void dTVResidCalc2D_kernel(float *Input1, float *Input2, float* Output, int N, int M, int num_total) +{ + int xIndex = blockDim.x * blockIdx.x + threadIdx.x; + int yIndex = blockDim.y * blockIdx.y + threadIdx.y; + + int index = xIndex + N*yIndex; + + if (index < num_total) { + Output[index] = Input1[index] - Input2[index]; + } +} + +__global__ void dTVResidCalc3D_kernel(float *Input1, float *Input2, float* Output, int N, int M, int Z, int num_total) +{ + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if (index < num_total) { + Output[index] = Input1[index] - Input2[index]; + } +} + +/************************************************/ +/*****************3D modules*********************/ +/************************************************/ +__global__ void GradNorm_func3D_kernel(float *Refd, float *Refd_x, float *Refd_y, float *Refd_z, float eta, int N, int M, int Z, int ImSize) +{ + + float val1, val2, val3, gradX, gradY, gradZ, magn; + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + /* boundary conditions */ + if (i >= N-1) val1 = 0.0f; else val1 = Refd[(N*M)*k + (i+1) + N*j]; + if (j >= M-1) val2 = 0.0f; else val2 = Refd[(N*M)*k + i + N*(j+1)]; + if (k >= Z-1) val3 = 0.0f; else val3 = Refd[(N*M)*(k+1) + i + N*j]; + + gradX = val1 - Refd[index]; + gradY = val2 - Refd[index]; + gradZ = val3 - Refd[index]; + magn = pow(gradX,2) + pow(gradY,2) + pow(gradZ,2); + magn = sqrt(magn + pow(eta,2)); + Refd_x[index] = gradX/magn; + Refd_y[index] = gradY/magn; + Refd_z[index] = gradZ/magn; + } + return; +} + +__global__ void ProjectVect_func3D_kernel(float *R1, float *R2, float *R3, float *Refd_x, float *Refd_y, float *Refd_z, int N, int M, int Z, int ImSize) +{ + + float in_prod; + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + in_prod = R1[index]*Refd_x[index] + R2[index]*Refd_y[index] + R3[index]*Refd_z[index]; /* calculate inner product */ + + R1[index] = R1[index] - in_prod*Refd_x[index]; + R2[index] = R2[index] - in_prod*Refd_y[index]; + R3[index] = R3[index] - in_prod*Refd_z[index]; + } + return; +} + + +__global__ void Obj_dfunc3D_kernel(float *Ad, float *D, float *R1, float *R2, float *R3, int N, int M, int Z, int ImSize, float lambda) +{ + + float val1,val2,val3; + + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + if (i <= 0) {val1 = 0.0f;} else {val1 = R1[(N*M)*(k) + (i-1) + N*j];} + if (j <= 0) {val2 = 0.0f;} else {val2 = R2[(N*M)*(k) + i + N*(j-1)];} + if (k <= 0) {val3 = 0.0f;} else {val3 = R3[(N*M)*(k-1) + i + N*j];} + //Write final result to global memory + D[index] = Ad[index] - lambda*(R1[index] + R2[index] + R3[index] - val1 - val2 - val3); + } + return; +} + +__global__ void Grad_dfunc3D_kernel(float *P1, float *P2, float *P3, float *D, float *R1, float *R2, float *R3, float *Refd_x, float *Refd_y, float *Refd_z, int N, int M, int Z, int ImSize, float multip) +{ + + float val1,val2,val3,in_prod; + + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + /* boundary conditions */ + if (i >= N-1) val1 = 0.0f; else val1 = D[index] - D[(N*M)*(k) + (i+1) + N*j]; + if (j >= M-1) val2 = 0.0f; else val2 = D[index] - D[(N*M)*(k) + i + N*(j+1)]; + if (k >= Z-1) val3 = 0.0f; else val3 = D[index] - D[(N*M)*(k+1) + i + N*j]; + + in_prod = val1*Refd_x[index] + val2*Refd_y[index] + val3*Refd_z[index]; /* calculate inner product */ + val1 = val1 - in_prod*Refd_x[index]; + val2 = val2 - in_prod*Refd_y[index]; + val3 = val3 - in_prod*Refd_z[index]; + + //Write final result to global memory + P1[index] = R1[index] + multip*val1; + P2[index] = R2[index] + multip*val2; + P3[index] = R3[index] + multip*val3; + } + return; +} + +__global__ void Proj_dfunc3D_iso_kernel(float *P1, float *P2, float *P3, int N, int M, int Z, int ImSize) +{ + + float denom,sq_denom; + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + denom = pow(P1[index],2) + pow(P2[index],2) + pow(P3[index],2); + + if (denom > 1.0f) { + sq_denom = 1.0f/sqrt(denom); + P1[index] = P1[index]*sq_denom; + P2[index] = P2[index]*sq_denom; + P3[index] = P3[index]*sq_denom; + } + } + return; +} + +__global__ void Proj_dfunc3D_aniso_kernel(float *P1, float *P2, float *P3, int N, int M, int Z, int ImSize) +{ + + float val1, val2, val3; + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + val1 = abs(P1[index]); + val2 = abs(P2[index]); + val3 = abs(P3[index]); + if (val1 < 1.0f) {val1 = 1.0f;} + if (val2 < 1.0f) {val2 = 1.0f;} + if (val3 < 1.0f) {val3 = 1.0f;} + P1[index] = P1[index]/val1; + P2[index] = P2[index]/val2; + P3[index] = P3[index]/val3; + } + return; +} + + +__global__ void Rupd_dfunc3D_kernel(float *P1, float *P1_old, float *P2, float *P2_old, float *P3, float *P3_old, float *R1, float *R2, float *R3, float tkp1, float tk, float multip2, int N, int M, int Z, int ImSize) +{ + //calculate each thread global index + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if ((i < N) && (j < M) && (k < Z)) { + R1[index] = P1[index] + multip2*(P1[index] - P1_old[index]); + R2[index] = P2[index] + multip2*(P2[index] - P2_old[index]); + R3[index] = P3[index] + multip2*(P3[index] - P3_old[index]); + } + return; +} + +__global__ void dTVnonneg3D_kernel(float* Output, int N, int M, int Z, int num_total) +{ + int i = blockDim.x * blockIdx.x + threadIdx.x; + int j = blockDim.y * blockIdx.y + threadIdx.y; + int k = blockDim.z * blockIdx.z + threadIdx.z; + + int index = (N*M)*k + i + N*j; + + if (index < num_total) { + if (Output[index] < 0.0f) Output[index] = 0.0f; + } +} +/*%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%*/ + +////////////MAIN HOST FUNCTION /////////////// +extern "C" int dTV_FGP_GPU_main(float *Input, float *InputRef, float *Output, float lambdaPar, int iter, float epsil, float eta, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ) +{ + int deviceCount = -1; // number of devices + cudaGetDeviceCount(&deviceCount); + if (deviceCount == 0) { + fprintf(stderr, "No CUDA devices found\n"); + return -1; + } + + int count = 0, i; + float re, multip,multip2; + float tk = 1.0f; + float tkp1=1.0f; + + if (dimZ <= 1) { + /*2D verson*/ + int ImSize = dimX*dimY; + float *d_input, *d_update=NULL, *d_update_prev=NULL, *P1=NULL, *P2=NULL, *P1_prev=NULL, *P2_prev=NULL, *R1=NULL, *R2=NULL, *InputRef_x=NULL, *InputRef_y=NULL, *d_InputRef=NULL; + + dim3 dimBlock(BLKXSIZE2D,BLKYSIZE2D); + dim3 dimGrid(idivup(dimX,BLKXSIZE2D), idivup(dimY,BLKYSIZE2D)); + + /*allocate space for images on device*/ + checkCudaErrors( cudaMalloc((void**)&d_input,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_update,ImSize*sizeof(float)) ); + if (epsil != 0.0f) checkCudaErrors( cudaMalloc((void**)&d_update_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P1,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P2,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P1_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P2_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R1,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R2,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_InputRef,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&InputRef_x,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&InputRef_y,ImSize*sizeof(float)) ); + + checkCudaErrors( cudaMemcpy(d_input,Input,ImSize*sizeof(float),cudaMemcpyHostToDevice)); + checkCudaErrors( cudaMemcpy(d_InputRef,InputRef,ImSize*sizeof(float),cudaMemcpyHostToDevice)); + + cudaMemset(P1, 0, ImSize*sizeof(float)); + cudaMemset(P2, 0, ImSize*sizeof(float)); + cudaMemset(P1_prev, 0, ImSize*sizeof(float)); + cudaMemset(P2_prev, 0, ImSize*sizeof(float)); + cudaMemset(R1, 0, ImSize*sizeof(float)); + cudaMemset(R2, 0, ImSize*sizeof(float)); + cudaMemset(InputRef_x, 0, ImSize*sizeof(float)); + cudaMemset(InputRef_y, 0, ImSize*sizeof(float)); + + /******************** Run CUDA 2D kernel here ********************/ + multip = (1.0f/(8.0f*lambdaPar)); + /* calculate gradient vectors for the reference */ + GradNorm_func2D_kernel<<<dimGrid,dimBlock>>>(d_InputRef, InputRef_x, InputRef_y, eta, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* The main kernel */ + for (i = 0; i < iter; i++) { + + /*projects a 2D vector field R-1,2 onto the orthogonal complement of another 2D vector field InputRef_xy*/ + ProjectVect_func2D_kernel<<<dimGrid,dimBlock>>>(R1, R2, InputRef_x, InputRef_y, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* computing the gradient of the objective function */ + Obj_dfunc2D_kernel<<<dimGrid,dimBlock>>>(d_input, d_update, R1, R2, dimX, dimY, ImSize, lambdaPar); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + if (nonneg != 0) { + dTVnonneg2D_kernel<<<dimGrid,dimBlock>>>(d_update, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); } + + /*Taking a step towards minus of the gradient*/ + Grad_dfunc2D_kernel<<<dimGrid,dimBlock>>>(P1, P2, d_update, R1, R2, InputRef_x, InputRef_y, dimX, dimY, ImSize, multip); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* projection step */ + if (methodTV == 0) Proj_dfunc2D_iso_kernel<<<dimGrid,dimBlock>>>(P1, P2, dimX, dimY, ImSize); /*isotropic TV*/ + else Proj_dfunc2D_aniso_kernel<<<dimGrid,dimBlock>>>(P1, P2, dimX, dimY, ImSize); /*anisotropic TV*/ + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + tkp1 = (1.0f + sqrt(1.0f + 4.0f*tk*tk))*0.5f; + multip2 = ((tk-1.0f)/tkp1); + + Rupd_dfunc2D_kernel<<<dimGrid,dimBlock>>>(P1, P1_prev, P2, P2_prev, R1, R2, tkp1, tk, multip2, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + if (epsil != 0.0f) { + /* calculate norm - stopping rules using the Thrust library */ + dTVResidCalc2D_kernel<<<dimGrid,dimBlock>>>(d_update, d_update_prev, P1_prev, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + thrust::device_vector<float> d_vec(P1_prev, P1_prev + ImSize); + float reduction = sqrt(thrust::transform_reduce(d_vec.begin(), d_vec.end(), square(), 0.0f, thrust::plus<float>())); + thrust::device_vector<float> d_vec2(d_update, d_update + ImSize); + float reduction2 = sqrt(thrust::transform_reduce(d_vec2.begin(), d_vec2.end(), square(), 0.0f, thrust::plus<float>())); + + re = (reduction/reduction2); + if (re < epsil) count++; + if (count > 4) break; + + dTVcopy_kernel2D<<<dimGrid,dimBlock>>>(d_update, d_update_prev, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + } + + dTVcopy_kernel2D<<<dimGrid,dimBlock>>>(P1, P1_prev, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + dTVcopy_kernel2D<<<dimGrid,dimBlock>>>(P2, P2_prev, dimX, dimY, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + tk = tkp1; + } + if (printM == 1) printf("FGP-dTV iterations stopped at iteration %i \n", i); + /***************************************************************/ + //copy result matrix from device to host memory + cudaMemcpy(Output,d_update,ImSize*sizeof(float),cudaMemcpyDeviceToHost); + + cudaFree(d_input); + cudaFree(d_update); + if (epsil != 0.0f) cudaFree(d_update_prev); + cudaFree(P1); + cudaFree(P2); + cudaFree(P1_prev); + cudaFree(P2_prev); + cudaFree(R1); + cudaFree(R2); + + cudaFree(d_InputRef); + cudaFree(InputRef_x); + cudaFree(InputRef_y); + } + else { + /*3D verson*/ + int ImSize = dimX*dimY*dimZ; + float *d_input, *d_update=NULL, *d_update_prev, *P1=NULL, *P2=NULL, *P3=NULL, *P1_prev=NULL, *P2_prev=NULL, *P3_prev=NULL, *R1=NULL, *R2=NULL, *R3=NULL, *InputRef_x=NULL, *InputRef_y=NULL, *InputRef_z=NULL, *d_InputRef=NULL; + + dim3 dimBlock(BLKXSIZE,BLKYSIZE,BLKZSIZE); + dim3 dimGrid(idivup(dimX,BLKXSIZE), idivup(dimY,BLKYSIZE),idivup(dimZ,BLKZSIZE)); + + /*allocate space for images on device*/ + checkCudaErrors( cudaMalloc((void**)&d_input,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_update,ImSize*sizeof(float)) ); + if (epsil != 0.0f) checkCudaErrors( cudaMalloc((void**)&d_update_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P1,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P2,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P3,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P1_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P2_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&P3_prev,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R1,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R2,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&R3,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&d_InputRef,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&InputRef_x,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&InputRef_y,ImSize*sizeof(float)) ); + checkCudaErrors( cudaMalloc((void**)&InputRef_z,ImSize*sizeof(float)) ); + + checkCudaErrors( cudaMemcpy(d_input,Input,ImSize*sizeof(float),cudaMemcpyHostToDevice)); + checkCudaErrors( cudaMemcpy(d_InputRef,InputRef,ImSize*sizeof(float),cudaMemcpyHostToDevice)); + + cudaMemset(P1, 0, ImSize*sizeof(float)); + cudaMemset(P2, 0, ImSize*sizeof(float)); + cudaMemset(P3, 0, ImSize*sizeof(float)); + cudaMemset(P1_prev, 0, ImSize*sizeof(float)); + cudaMemset(P2_prev, 0, ImSize*sizeof(float)); + cudaMemset(P3_prev, 0, ImSize*sizeof(float)); + cudaMemset(R1, 0, ImSize*sizeof(float)); + cudaMemset(R2, 0, ImSize*sizeof(float)); + cudaMemset(R3, 0, ImSize*sizeof(float)); + cudaMemset(InputRef_x, 0, ImSize*sizeof(float)); + cudaMemset(InputRef_y, 0, ImSize*sizeof(float)); + cudaMemset(InputRef_z, 0, ImSize*sizeof(float)); + + /********************** Run CUDA 3D kernel here ********************/ + multip = (1.0f/(26.0f*lambdaPar)); + /* calculate gradient vectors for the reference */ + GradNorm_func3D_kernel<<<dimGrid,dimBlock>>>(d_InputRef, InputRef_x, InputRef_y, InputRef_z, eta, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* The main kernel */ + for (i = 0; i < iter; i++) { + + /*projects a 3D vector field R-1,2,3 onto the orthogonal complement of another 3D vector field InputRef_xyz*/ + ProjectVect_func3D_kernel<<<dimGrid,dimBlock>>>(R1, R2, R3, InputRef_x, InputRef_y, InputRef_z, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* computing the gradient of the objective function */ + Obj_dfunc3D_kernel<<<dimGrid,dimBlock>>>(d_input, d_update, R1, R2, R3, dimX, dimY, dimZ, ImSize, lambdaPar); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + if (nonneg != 0) { + dTVnonneg3D_kernel<<<dimGrid,dimBlock>>>(d_update, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); } + + /*Taking a step towards minus of the gradient*/ + Grad_dfunc3D_kernel<<<dimGrid,dimBlock>>>(P1, P2, P3, d_update, R1, R2, R3, InputRef_x, InputRef_y, InputRef_z, dimX, dimY, dimZ, ImSize, multip); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + /* projection step */ + if (methodTV == 0) Proj_dfunc3D_iso_kernel<<<dimGrid,dimBlock>>>(P1, P2, P3, dimX, dimY, dimZ, ImSize); /* isotropic kernel */ + else Proj_dfunc3D_aniso_kernel<<<dimGrid,dimBlock>>>(P1, P2, P3, dimX, dimY, dimZ, ImSize); /* anisotropic kernel */ + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + tkp1 = (1.0f + sqrt(1.0f + 4.0f*tk*tk))*0.5f; + multip2 = ((tk-1.0f)/tkp1); + + Rupd_dfunc3D_kernel<<<dimGrid,dimBlock>>>(P1, P1_prev, P2, P2_prev, P3, P3_prev, R1, R2, R3, tkp1, tk, multip2, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + if (epsil != 0.0f) { + /* calculate norm - stopping rules using the Thrust library */ + dTVResidCalc3D_kernel<<<dimGrid,dimBlock>>>(d_update, d_update_prev, P1_prev, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + thrust::device_vector<float> d_vec(P1_prev, P1_prev + ImSize); + float reduction = sqrt(thrust::transform_reduce(d_vec.begin(), d_vec.end(), square(), 0.0f, thrust::plus<float>())); + thrust::device_vector<float> d_vec2(d_update, d_update + ImSize); + float reduction2 = sqrt(thrust::transform_reduce(d_vec2.begin(), d_vec2.end(), square(), 0.0f, thrust::plus<float>())); + + re = (reduction/reduction2); + if (re < epsil) count++; + if (count > 4) break; + + dTVcopy_kernel3D<<<dimGrid,dimBlock>>>(d_update, d_update_prev, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + } + + dTVcopy_kernel3D<<<dimGrid,dimBlock>>>(P1, P1_prev, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + dTVcopy_kernel3D<<<dimGrid,dimBlock>>>(P2, P2_prev, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + dTVcopy_kernel3D<<<dimGrid,dimBlock>>>(P3, P3_prev, dimX, dimY, dimZ, ImSize); + checkCudaErrors( cudaDeviceSynchronize() ); + checkCudaErrors(cudaPeekAtLastError() ); + + tk = tkp1; + } + if (printM == 1) printf("FGP-dTV iterations stopped at iteration %i \n", i); + /***************************************************************/ + //copy result matrix from device to host memory + cudaMemcpy(Output,d_update,ImSize*sizeof(float),cudaMemcpyDeviceToHost); + + cudaFree(d_input); + cudaFree(d_update); + if (epsil != 0.0f) cudaFree(d_update_prev); + cudaFree(P1); + cudaFree(P2); + cudaFree(P3); + cudaFree(P1_prev); + cudaFree(P2_prev); + cudaFree(P3_prev); + cudaFree(R1); + cudaFree(R2); + cudaFree(R3); + cudaFree(InputRef_x); + cudaFree(InputRef_y); + cudaFree(InputRef_z); + cudaFree(d_InputRef); + } + //cudaDeviceReset(); + return 0; +} diff --git a/src/Core/regularisers_GPU/dTV_FGP_GPU_core.h b/src/Core/regularisers_GPU/dTV_FGP_GPU_core.h new file mode 100644 index 0000000..f9281e8 --- /dev/null +++ b/src/Core/regularisers_GPU/dTV_FGP_GPU_core.h @@ -0,0 +1,9 @@ +#ifndef _dTV_FGP_GPU_ +#define _dTV_FGP_GPU_ + +#include "CCPiDefines.h" +#include <memory.h> + +extern "C" CCPI_EXPORT int dTV_FGP_GPU_main(float *Input, float *InputRef, float *Output, float lambdaPar, int iter, float epsil, float eta, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ); + +#endif diff --git a/src/Core/regularisers_GPU/shared.h b/src/Core/regularisers_GPU/shared.h new file mode 100644 index 0000000..fe98cd6 --- /dev/null +++ b/src/Core/regularisers_GPU/shared.h @@ -0,0 +1,42 @@ +/*shared macros*/ + + +/*checks CUDA call, should be used in functions returning <int> value +if error happens, writes to standard error and explicitly returns -1*/ +#define CHECK(call) \ +{ \ + const cudaError_t error = call; \ + if (error != cudaSuccess) \ + { \ + fprintf(stderr, "Error: %s:%d, ", __FILE__, __LINE__); \ + fprintf(stderr, "code: %d, reason: %s\n", error, \ + cudaGetErrorString(error)); \ + return -1; \ + } \ +} + +// This will output the proper CUDA error strings in the event that a CUDA host call returns an error +#define checkCudaErrors(call) \ +{ \ + const cudaError_t error = call; \ + if (error != cudaSuccess) \ + { \ + fprintf(stderr, "Error: %s:%d, ", __FILE__, __LINE__); \ + fprintf(stderr, "code: %d, reason: %s\n", error, \ + cudaGetErrorString(error)); \ + return -1; \ + } \ +} +/*#define checkCudaErrors(err) __checkCudaErrors (err, __FILE__, __LINE__) + +inline void __checkCudaErrors(cudaError err, const char *file, const int line) +{ + if (cudaSuccess != err) + { + fprintf(stderr, "%s(%i) : CUDA Runtime API error %d: %s.\n", + file, line, (int)err, cudaGetErrorString(err)); + return; + } +} +*/ + diff --git a/src/Matlab/CMakeLists.txt b/src/Matlab/CMakeLists.txt new file mode 100755 index 0000000..b97f845 --- /dev/null +++ b/src/Matlab/CMakeLists.txt @@ -0,0 +1,147 @@ +project(regulariserMatlab)
+
+
+find_package(Matlab REQUIRED COMPONENTS MAIN_PROGRAM MX_LIBRARY ENG_LIBRARY )
+
+
+
+#C:\Users\ofn77899\Documents\Projects\CCPi\GitHub\CCPi-FISTA_Reconstruction\Core\regularisers_CPU
+# matlab_add_mex(
+ # NAME CPU_ROF
+ # SRC
+ # ${CMAKE_SOURCE_DIR}/Matlab/mex_compile/regularisers_CPU/ROF_TV.c
+ # LINK_TO cilreg ${Matlab_LIBRARIES}
+ # )
+
+# target_include_directories(CPU_ROF
+ # PUBLIC ${CMAKE_SOURCE_DIR}/Core/regularisers_CPU
+ # ${CMAKE_SOURCE_DIR}/Core/regularisers_GPU
+ # ${CMAKE_SOURCE_DIR}/Core/inpainters_CPU
+ # ${CMAKE_SOURCE_DIR}/Core/
+ # ${MATLAB_INCLUDE_DIR})
+
+ # matlab_add_mex(
+ # NAME CPU_TNV
+ # SRC
+ # ${CMAKE_SOURCE_DIR}/Matlab/mex_compile/regularisers_CPU/TNV.c
+ # LINK_TO cilreg ${Matlab_LIBRARIES}
+ # )
+
+# target_include_directories(CPU_TNV
+ # PUBLIC ${CMAKE_SOURCE_DIR}/Core/regularisers_CPU
+ # ${CMAKE_SOURCE_DIR}/Core/regularisers_GPU
+ # ${CMAKE_SOURCE_DIR}/Core/inpainters_CPU
+ # ${CMAKE_SOURCE_DIR}/Core/
+ # ${MATLAB_INCLUDE_DIR})
+
+#set (CPU_MEX_FILES "regularisers_CPU/TNV.c;regularisers_CPU/ROF_TV.c")
+#set (MEX_TARGETS "CPU_TNV;CPU_ROF")
+#list(APPEND MEX_TARGETS "CPU_TNV")
+#list(APPEND MEX_TARGETS "CPU_ROF")
+
+file(GLOB CPU_MEX_FILES
+ "${CMAKE_SOURCE_DIR}/Matlab/mex_compile/regularisers_CPU/*.c"
+ #"${CMAKE_SOURCE_DIR}/Matlab/mex_compile/regularisers_GPU/*.c"
+)
+
+#message("CPU_MEX_FILES " ${CPU_MEX_FILES})
+
+list(LENGTH CPU_MEX_FILES num)
+
+
+MATH(EXPR num "${num}-1")
+#set(num "-1")
+message("found ${num} files")
+
+foreach(tgt RANGE 0 ${num})
+ message("number " ${tgt})
+ list(LENGTH CPU_MEX_FILES num2)
+ message("the list is ${num2}")
+ #list(GET CPU_TARGETS ${tgt} current_target)
+ list(GET CPU_MEX_FILES ${tgt} current_file_name)
+ get_filename_component(current_file ${current_file_name} NAME)
+ string(REGEX MATCH "(.+).c" match ${current_file})
+ if (NOT ${match} EQUAL "" )
+ set (current_target ${CMAKE_MATCH_1})
+ endif()
+ message("matlab_add_mex target " ${current_file} " and " ${current_target})
+ matlab_add_mex(
+ NAME ${current_target}
+ SRC
+ ${current_file_name}
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/FGP_TV_core.c
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/SB_TV_core.c
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/TGV_core.c
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/Diffusion_core.c
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/Diffus4th_order_core.c
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/LLT_ROF_core.c
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/ROF_TV_core.c
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/FGP_dTV_core.c
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/TNV_core.c
+ #${CMAKE_SOURCE_DIR}/Core/regularisers_CPU/utils.c
+ #${CMAKE_SOURCE_DIR}/Core/inpainters_CPU/Diffusion_Inpaint_core.c
+ #${CMAKE_SOURCE_DIR}/Core/inpainters_CPU/NonlocalMarching_Inpaint_core.c
+ LINK_TO cilreg ${Matlab_LIBRARIES}
+ )
+
+target_include_directories(${current_target}
+ PUBLIC ${CMAKE_SOURCE_DIR}/Core/regularisers_CPU
+ ${CMAKE_SOURCE_DIR}/Core/regularisers_GPU
+ ${CMAKE_SOURCE_DIR}/Core/inpainters_CPU
+ ${CMAKE_SOURCE_DIR}/Core/
+ ${MATLAB_INCLUDE_DIR})
+ set_property(TARGET ${current_target} PROPERTY C_STANDARD 99)
+ list(APPEND CPU_MEX_TARGETS ${current_target})
+ INSTALL(TARGETS ${current_target} DESTINATION "${MATLAB_DEST}")
+endforeach()
+
+add_custom_target(MatlabWrapper DEPENDS ${CPU_MEX_TARGETS})
+
+if (BUILD_CUDA)
+ find_package(CUDA)
+ if (CUDA_FOUND)
+ file(GLOB GPU_MEX_FILES
+ "${CMAKE_SOURCE_DIR}/Matlab/mex_compile/regularisers_GPU/*.cpp"
+ )
+
+ list(LENGTH GPU_MEX_FILES num)
+message("number of GPU files " ${num})
+
+ MATH(EXPR num "${num}-1")
+ #set(num "-1")
+
+ foreach(tgt RANGE ${num})
+ message("number " ${tgt})
+ list(LENGTH GPU_MEX_FILES num2)
+ message("the list is ${num2}")
+ #list(GET CPU_TARGETS ${tgt} current_target)
+ list(GET GPU_MEX_FILES ${tgt} current_file_name)
+ get_filename_component(current_file ${current_file_name} NAME)
+ string(REGEX MATCH "(.+).c" match ${current_file})
+ if (NOT ${match} EQUAL "" )
+ set (current_target ${CMAKE_MATCH_1})
+ endif()
+ message("matlab_add_mex target " ${current_file} " and " ${current_target})
+ message("matlab_add_mex " ${current_target})
+ matlab_add_mex(
+ NAME ${current_target}
+ SRC
+ ${current_file_name}
+ LINK_TO cilregcuda ${Matlab_LIBRARIES}
+ )
+
+ target_include_directories(${current_target}
+ PUBLIC ${CMAKE_SOURCE_DIR}/Core/regularisers_CPU
+ ${CMAKE_SOURCE_DIR}/Core/regularisers_GPU
+ ${CMAKE_SOURCE_DIR}/Core/inpainters_CPU
+ ${CMAKE_SOURCE_DIR}/Core/
+ ${MATLAB_INCLUDE_DIR})
+
+ list(APPEND GPU_MEX_TARGETS ${current_target})
+ INSTALL(TARGETS ${current_target} DESTINATION "${MATLAB_DEST}")
+ endforeach()
+
+ add_custom_target(MatlabWrapperGPU DEPENDS ${GPU_MEX_TARGETS})
+
+ endif()
+endif()
diff --git a/src/Matlab/mex_compile/compileCPU_mex_Linux.m b/src/Matlab/mex_compile/compileCPU_mex_Linux.m new file mode 100644 index 0000000..72a828e --- /dev/null +++ b/src/Matlab/mex_compile/compileCPU_mex_Linux.m @@ -0,0 +1,81 @@ +% execute this mex file on Linux in Matlab once + +fsep = '/'; + +pathcopyFrom = sprintf(['..' fsep '..' fsep '..' fsep 'Core' fsep 'regularisers_CPU'], 1i); +pathcopyFrom1 = sprintf(['..' fsep '..' fsep '..' fsep 'Core' fsep 'CCPiDefines.h'], 1i); +pathcopyFrom2 = sprintf(['..' fsep '..' fsep '..' fsep 'Core' fsep 'inpainters_CPU'], 1i); + +copyfile(pathcopyFrom, 'regularisers_CPU'); +copyfile(pathcopyFrom1, 'regularisers_CPU'); +copyfile(pathcopyFrom2, 'regularisers_CPU'); + +cd regularisers_CPU + +Pathmove = sprintf(['..' fsep 'installed' fsep], 1i); + +fprintf('%s \n', '<<<<<<<<<<<Compiling CPU regularisers>>>>>>>>>>>>>'); + +fprintf('%s \n', 'Compiling ROF-TV...'); +mex ROF_TV.c ROF_TV_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('ROF_TV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling FGP-TV...'); +mex FGP_TV.c FGP_TV_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('FGP_TV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling SB-TV...'); +mex SB_TV.c SB_TV_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('SB_TV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling dFGP-TV...'); +mex FGP_dTV.c FGP_dTV_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('FGP_dTV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling TNV...'); +mex TNV.c TNV_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('TNV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling NonLinear Diffusion...'); +mex NonlDiff.c Diffusion_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('NonlDiff.mex*',Pathmove); + +fprintf('%s \n', 'Compiling Anisotropic diffusion of higher order...'); +mex Diffusion_4thO.c Diffus4th_order_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('Diffusion_4thO.mex*',Pathmove); + +fprintf('%s \n', 'Compiling TGV...'); +mex TGV.c TGV_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('TGV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling ROF-LLT...'); +mex LLT_ROF.c LLT_ROF_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('LLT_ROF.mex*',Pathmove); + +fprintf('%s \n', 'Compiling NonLocal-TV...'); +mex PatchSelect.c PatchSelect_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +mex Nonlocal_TV.c Nonlocal_TV_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('Nonlocal_TV.mex*',Pathmove); +movefile('PatchSelect.mex*',Pathmove); + +fprintf('%s \n', 'Compiling additional tools...'); +mex TV_energy.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('TV_energy.mex*',Pathmove); + +%############Inpainters##############% +fprintf('%s \n', 'Compiling Nonlinear/Linear diffusion inpainting...'); +mex NonlDiff_Inp.c Diffusion_Inpaint_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('NonlDiff_Inp.mex*',Pathmove); + +fprintf('%s \n', 'Compiling Nonlocal marching method for inpainting...'); +mex NonlocalMarching_Inpaint.c NonlocalMarching_Inpaint_core.c utils.c CFLAGS="\$CFLAGS -fopenmp -Wall -std=c99" LDFLAGS="\$LDFLAGS -fopenmp" +movefile('NonlocalMarching_Inpaint.mex*',Pathmove); + +delete SB_TV_core* ROF_TV_core* FGP_TV_core* FGP_dTV_core* TNV_core* utils* Diffusion_core* Diffus4th_order_core* TGV_core* LLT_ROF_core* CCPiDefines.h +delete PatchSelect_core* Nonlocal_TV_core* +delete Diffusion_Inpaint_core* NonlocalMarching_Inpaint_core* +fprintf('%s \n', '<<<<<<< Regularisers successfully compiled! >>>>>>>'); + +pathA2 = sprintf(['..' fsep '..' fsep], 1i); +cd(pathA2); +cd demos diff --git a/src/Matlab/mex_compile/compileCPU_mex_WINDOWS.m b/src/Matlab/mex_compile/compileCPU_mex_WINDOWS.m new file mode 100644 index 0000000..6f7541c --- /dev/null +++ b/src/Matlab/mex_compile/compileCPU_mex_WINDOWS.m @@ -0,0 +1,135 @@ +% execute this mex file on Windows in Matlab once + +% >>>>>>>>>>>>>>>>>>>>>>>>>>>>> +% I've been able to compile on Windows 7 with MinGW and Matlab 2016b, however, +% not sure if openmp is enabled after the compilation. + +% Here I present two ways how software can be compiled, if you have some +% other suggestions/remarks please contact me at dkazanc@hotmail.com +% >>>>>>>>>>>>>>>>>>>>>>>>>>>>> + +fsep = '/'; + +pathcopyFrom = sprintf(['..' fsep '..' fsep '..' fsep 'Core' fsep 'regularisers_CPU'], 1i); +pathcopyFrom1 = sprintf(['..' fsep '..' fsep '..' fsep 'Core' fsep 'CCPiDefines.h'], 1i); +pathcopyFrom2 = sprintf(['..' fsep '..' fsep '..' fsep 'Core' fsep 'inpainters_CPU'], 1i); + +copyfile(pathcopyFrom, 'regularisers_CPU'); +copyfile(pathcopyFrom1, 'regularisers_CPU'); +copyfile(pathcopyFrom2, 'regularisers_CPU'); + +cd regularisers_CPU + +Pathmove = sprintf(['..' fsep 'installed' fsep], 1i); + +fprintf('%s \n', '<<<<<<<<<<<Compiling CPU regularisers>>>>>>>>>>>>>'); + +fprintf('%s \n', 'Compiling ROF-TV...'); +mex ROF_TV.c ROF_TV_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('ROF_TV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling FGP-TV...'); +mex FGP_TV.c FGP_TV_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('FGP_TV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling SB-TV...'); +mex SB_TV.c SB_TV_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('SB_TV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling dFGP-TV...'); +mex FGP_dTV.c FGP_dTV_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('FGP_dTV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling TNV...'); +mex TNV.c TNV_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('TNV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling NonLinear Diffusion...'); +mex NonlDiff.c Diffusion_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('NonlDiff.mex*',Pathmove); + +fprintf('%s \n', 'Compiling Anisotropic diffusion of higher order...'); +mex Diffusion_4thO.c Diffus4th_order_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('Diffusion_4thO.mex*',Pathmove); + +fprintf('%s \n', 'Compiling TGV...'); +mex TGV.c TGV_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('TGV.mex*',Pathmove); + +fprintf('%s \n', 'Compiling ROF-LLT...'); +mex LLT_ROF.c LLT_ROF_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('LLT_ROF.mex*',Pathmove); + +fprintf('%s \n', 'Compiling NonLocal-TV...'); +mex PatchSelect.c PatchSelect_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +mex Nonlocal_TV.c Nonlocal_TV_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('Nonlocal_TV.mex*',Pathmove); +movefile('PatchSelect.mex*',Pathmove); + +fprintf('%s \n', 'Compiling additional tools...'); +mex TV_energy.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('TV_energy.mex*',Pathmove); + +%############Inpainters##############% +fprintf('%s \n', 'Compiling Nonlinear/Linear diffusion inpainting...'); +mex NonlDiff_Inp.c Diffusion_Inpaint_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('NonlDiff_Inp.mex*',Pathmove); + +fprintf('%s \n', 'Compiling Nonlocal marching method for inpaiting...'); +mex NonlocalMarching_Inpaint.c NonlocalMarching_Inpaint_core.c utils.c COMPFLAGS="\$COMPFLAGS -fopenmp -Wall -std=c99" +movefile('NonlocalMarching_Inpaint.mex*',Pathmove); + + +%% +%%% The second approach to compile using TDM-GCC which follows this +%%% discussion: +%%% https://uk.mathworks.com/matlabcentral/answers/279171-using-mingw-compiler-and-open-mp#comment_359122 +%%% 1. Install TDM-GCC independently from http://tdm-gcc.tdragon.net/ (I installed 5.1.0) +%%% Install openmp version: http://sourceforge.net/projects/tdm-gcc/files/TDM-GCC%205%20series/5.1.0-tdm64-1/gcc-5.1.0-tdm64-1-openmp.zip/download +%%% 2. Link til libgomp.a in that installation when compilling your mex file. + +%%% assuming you unzipped TDM GCC (OpenMp) in folder TDMGCC on C drive, uncomment +%%% bellow +% fprintf('%s \n', 'Compiling CPU regularisers...'); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" ROF_TV.c ROF_TV_core.c utils.c +% movefile('ROF_TV.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" FGP_TV.c FGP_TV_core.c utils.c +% movefile('FGP_TV.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" SB_TV.c SB_TV_core.c utils.c +% movefile('SB_TV.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" FGP_dTV.c FGP_dTV_core.c utils.c +% movefile('FGP_dTV.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" TNV.c TNV_core.c utils.c +% movefile('TNV.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" NonlDiff.c Diffusion_core.c utils.c +% movefile('NonlDiff.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" Diffusion_4thO.c Diffus4th_order_core.c utils.c +% movefile('Diffusion_4thO.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" TGV.c TGV_core.c utils.c +% movefile('TGV.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" LLT_ROF.c LLT_ROF_core.c utils.c +% movefile('LLT_ROF.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" PatchSelect.c PatchSelect_core.c utils.c +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" Nonlocal_TV.c Nonlocal_TV_core.c utils.c +% movefile('Nonlocal_TV.mex*',Pathmove); +% movefile('PatchSelect.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" TV_energy.c utils.c +% movefile('TV_energy.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" NonlDiff_Inp.c Diffusion_Inpaint_core.c utils.c +% movefile('NonlDiff_Inp.mex*',Pathmove); +% mex C:\TDMGCC\lib\gcc\x86_64-w64-mingw32\5.1.0\libgomp.a CXXFLAGS="$CXXFLAGS -std=c++11 -fopenmp" NonlocalMarching_Inpaint.c NonlocalMarching_Inpaint_core.c utils.c +% movefile('NonlocalMarching_Inpaint.mex*',Pathmove); + + +delete SB_TV_core* ROF_TV_core* FGP_TV_core* FGP_dTV_core* TNV_core* utils* Diffusion_core* Diffus4th_order_core* TGV_core* CCPiDefines.h +delete PatchSelect_core* Nonlocal_TV_core* +delete Diffusion_Inpaint_core* NonlocalMarching_Inpaint_core* +fprintf('%s \n', 'Regularisers successfully compiled!'); + + +%% +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + +%pathA2 = sprintf(['..' fsep '..' fsep], 1i); +%cd(pathA2); +%cd demos diff --git a/src/Matlab/mex_compile/compileGPU_mex.m b/src/Matlab/mex_compile/compileGPU_mex.m new file mode 100644 index 0000000..dd1475c --- /dev/null +++ b/src/Matlab/mex_compile/compileGPU_mex.m @@ -0,0 +1,74 @@ +% execute this mex file in Matlab once + +%>>>>>>>>>>>>>>>>>Important<<<<<<<<<<<<<<<<<<< +% In order to compile CUDA modules one needs to have nvcc-compiler +% installed (see CUDA SDK), check it under MATLAB with !nvcc --version + +% In the code bellow we provide a full explicit path to nvcc compiler +% ! paths to matlab and CUDA sdk can be different, modify accordingly ! + +% Tested on Ubuntu 18.04/MATLAB 2016b/cuda10.0/gcc7.3 + +% Installation HAS NOT been tested on Windows, please you Cmake build or +% modify the code bellow accordingly +fsep = '/'; + +pathcopyFrom = sprintf(['..' fsep '..' fsep '..' fsep 'Core' fsep 'regularisers_GPU'], 1i); +pathcopyFrom1 = sprintf(['..' fsep '..' fsep '..' fsep 'Core' fsep 'CCPiDefines.h'], 1i); + +copyfile(pathcopyFrom, 'regularisers_GPU'); +copyfile(pathcopyFrom1, 'regularisers_GPU'); + +cd regularisers_GPU + +Pathmove = sprintf(['..' fsep 'installed' fsep], 1i); + +fprintf('%s \n', '<<<<<<<<<<<Compiling GPU regularisers (CUDA)>>>>>>>>>>>>>'); + +fprintf('%s \n', 'Compiling ROF-TV...'); +!/usr/local/cuda/bin/nvcc -O0 -c TV_ROF_GPU_core.cu -Xcompiler -fPIC -I~/SOFT/MATLAB9/extern/include/ +mex -g -I/usr/local/cuda-10.0/include -L/usr/local/cuda-10.0/lib64 -lcudart -lcufft -lmwgpu ROF_TV_GPU.cpp TV_ROF_GPU_core.o +movefile('ROF_TV_GPU.mex*',Pathmove); + +fprintf('%s \n', 'Compiling FGP-TV...'); +!/usr/local/cuda/bin/nvcc -O0 -c TV_FGP_GPU_core.cu -Xcompiler -fPIC -I~/SOFT/MATLAB9/extern/include/ +mex -g -I/usr/local/cuda-10.0/include -L/usr/local/cuda-10.0/lib64 -lcudart -lcufft -lmwgpu FGP_TV_GPU.cpp TV_FGP_GPU_core.o +movefile('FGP_TV_GPU.mex*',Pathmove); + +fprintf('%s \n', 'Compiling SB-TV...'); +!/usr/local/cuda/bin/nvcc -O0 -c TV_SB_GPU_core.cu -Xcompiler -fPIC -I~/SOFT/MATLAB9/extern/include/ +mex -g -I/usr/local/cuda-10.0/include -L/usr/local/cuda-10.0/lib64 -lcudart -lcufft -lmwgpu SB_TV_GPU.cpp TV_SB_GPU_core.o +movefile('SB_TV_GPU.mex*',Pathmove); + +fprintf('%s \n', 'Compiling TGV...'); +!/usr/local/cuda/bin/nvcc -O0 -c TGV_GPU_core.cu -Xcompiler -fPIC -I~/SOFT/MATLAB9/extern/include/ +mex -g -I/usr/local/cuda-10.0/include -L/usr/local/cuda-10.0/lib64 -lcudart -lcufft -lmwgpu TGV_GPU.cpp TGV_GPU_core.o +movefile('TGV_GPU.mex*',Pathmove); + +fprintf('%s \n', 'Compiling dFGP-TV...'); +!/usr/local/cuda/bin/nvcc -O0 -c dTV_FGP_GPU_core.cu -Xcompiler -fPIC -I~/SOFT/MATLAB9/extern/include/ +mex -g -I/usr/local/cuda-10.0/include -L/usr/local/cuda-10.0/lib64 -lcudart -lcufft -lmwgpu FGP_dTV_GPU.cpp dTV_FGP_GPU_core.o +movefile('FGP_dTV_GPU.mex*',Pathmove); + +fprintf('%s \n', 'Compiling NonLinear Diffusion...'); +!/usr/local/cuda/bin/nvcc -O0 -c NonlDiff_GPU_core.cu -Xcompiler -fPIC -I~/SOFT/MATLAB9/extern/include/ +mex -g -I/usr/local/cuda-10.0/include -L/usr/local/cuda-10.0/lib64 -lcudart -lcufft -lmwgpu NonlDiff_GPU.cpp NonlDiff_GPU_core.o +movefile('NonlDiff_GPU.mex*',Pathmove); + +fprintf('%s \n', 'Compiling Anisotropic diffusion of higher order...'); +!/usr/local/cuda/bin/nvcc -O0 -c Diffus_4thO_GPU_core.cu -Xcompiler -fPIC -I~/SOFT/MATLAB9/extern/include/ +mex -g -I/usr/local/cuda-10.0/include -L/usr/local/cuda-10.0/lib64 -lcudart -lcufft -lmwgpu Diffusion_4thO_GPU.cpp Diffus_4thO_GPU_core.o +movefile('Diffusion_4thO_GPU.mex*',Pathmove); + +fprintf('%s \n', 'Compiling ROF-LLT...'); +!/usr/local/cuda/bin/nvcc -O0 -c LLT_ROF_GPU_core.cu -Xcompiler -fPIC -I~/SOFT/MATLAB9/extern/include/ +mex -g -I/usr/local/cuda-10.0/include -L/usr/local/cuda-10.0/lib64 -lcudart -lcufft -lmwgpu LLT_ROF_GPU.cpp LLT_ROF_GPU_core.o +movefile('LLT_ROF_GPU.mex*',Pathmove); + + +delete TV_ROF_GPU_core* TV_FGP_GPU_core* TV_SB_GPU_core* dTV_FGP_GPU_core* NonlDiff_GPU_core* Diffus_4thO_GPU_core* TGV_GPU_core* LLT_ROF_GPU_core* CCPiDefines.h +fprintf('%s \n', 'All successfully compiled!'); + +pathA2 = sprintf(['..' fsep '..' fsep], 1i); +cd(pathA2); +cd demos
\ No newline at end of file diff --git a/src/Matlab/mex_compile/installed/MEXed_files_location.txt b/src/Matlab/mex_compile/installed/MEXed_files_location.txt new file mode 100644 index 0000000..e69de29 --- /dev/null +++ b/src/Matlab/mex_compile/installed/MEXed_files_location.txt diff --git a/src/Matlab/mex_compile/regularisers_CPU/Diffusion_4thO.c b/src/Matlab/mex_compile/regularisers_CPU/Diffusion_4thO.c new file mode 100644 index 0000000..66ea9be --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/Diffusion_4thO.c @@ -0,0 +1,77 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "Diffus4th_order_core.h" + +/* C-OMP implementation of fourth-order diffusion scheme [1] for piecewise-smooth recovery (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. lambda - regularization parameter [REQUIRED] + * 3. Edge-preserving parameter (sigma) [REQUIRED] + * 4. Number of iterations, for explicit scheme >= 150 is recommended [OPTIONAL, default 300] + * 5. tau - time-marching step for the explicit scheme [OPTIONAL, default 0.015] + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Hajiaboli, M.R., 2011. An anisotropic fourth-order diffusion filter for image noise removal. International Journal of Computer Vision, 92(2), pp.177-191. + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter_numb; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + float *Input, *Output=NULL, lambda, tau, sigma; + + dim_array = mxGetDimensions(prhs[0]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + sigma = (float) mxGetScalar(prhs[2]); /* Edge-preserving parameter */ + iter_numb = 300; /* iterations number */ + tau = 0.01; /* marching step parameter */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if ((nrhs < 3) || (nrhs > 5)) mexErrMsgTxt("At least 3 parameters is required, all parameters are: Image(2D/3D), Regularisation parameter, Edge-preserving parameter, iterations number, time-marching constant"); + if ((nrhs == 4) || (nrhs == 5)) iter_numb = (int) mxGetScalar(prhs[3]); /* iterations number */ + if (nrhs == 5) tau = (float) mxGetScalar(prhs[4]); /* marching step parameter */ + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + Diffus4th_CPU_main(Input, Output, lambda, sigma, iter_numb, tau, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_CPU/FGP_TV.c b/src/Matlab/mex_compile/regularisers_CPU/FGP_TV.c new file mode 100644 index 0000000..642362f --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/FGP_TV.c @@ -0,0 +1,97 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "FGP_TV_core.h" + +/* C-OMP implementation of FGP-TV [1] denoising/regularization model (2D/3D case) + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambdaPar - regularization parameter + * 3. Number of iterations + * 4. eplsilon: tolerance constant + * 5. TV-type: methodTV - 'iso' (0) or 'l1' (1) + * 6. nonneg: 'nonnegativity (0 is OFF by default) + * 7. print information: 0 (off) or 1 (on) + * + * Output: + * [1] Filtered/regularized image + * + * This function is based on the Matlab's code and paper by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + */ + + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter, methTV, printswitch, nonneg; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + float *Input, *Output=NULL, lambda, epsil; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + + /*Handling Matlab input data*/ + if ((nrhs < 2) || (nrhs > 7)) mexErrMsgTxt("At least 2 parameters is required, all parameters are: Image(2D/3D), Regularization parameter, Regularization parameter, iterations number, tolerance, penalty type ('iso' or 'l1'), nonnegativity switch, print switch"); + + Input = (float *) mxGetData(prhs[0]); /*noisy image (2D/3D) */ + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + iter = 300; /* default iterations number */ + epsil = 0.0001; /* default tolerance constant */ + methTV = 0; /* default isotropic TV penalty */ + nonneg = 0; /* default nonnegativity switch, off - 0 */ + printswitch = 0; /*default print is switched, off - 0 */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + + if ((nrhs == 3) || (nrhs == 4) || (nrhs == 5) || (nrhs == 6) || (nrhs == 7)) iter = (int) mxGetScalar(prhs[2]); /* iterations number */ + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6) || (nrhs == 7)) epsil = (float) mxGetScalar(prhs[3]); /* tolerance constant */ + if ((nrhs == 5) || (nrhs == 6) || (nrhs == 7)) { + char *penalty_type; + penalty_type = mxArrayToString(prhs[4]); /* choosing TV penalty: 'iso' or 'l1', 'iso' is the default */ + if ((strcmp(penalty_type, "l1") != 0) && (strcmp(penalty_type, "iso") != 0)) mexErrMsgTxt("Choose TV type: 'iso' or 'l1',"); + if (strcmp(penalty_type, "l1") == 0) methTV = 1; /* enable 'l1' penalty */ + mxFree(penalty_type); + } + if ((nrhs == 6) || (nrhs == 7)) { + nonneg = (int) mxGetScalar(prhs[5]); + if ((nonneg != 0) && (nonneg != 1)) mexErrMsgTxt("Nonnegativity constraint can be enabled by choosing 1 or off - 0"); + } + if (nrhs == 7) { + printswitch = (int) mxGetScalar(prhs[6]); + if ((printswitch != 0) && (printswitch != 1)) mexErrMsgTxt("Print can be enabled by choosing 1 or off - 0"); + } + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + /* running the function */ + TV_FGP_CPU_main(Input, Output, lambda, iter, epsil, methTV, nonneg, printswitch, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_CPU/FGP_dTV.c b/src/Matlab/mex_compile/regularisers_CPU/FGP_dTV.c new file mode 100644 index 0000000..1a0c070 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/FGP_dTV.c @@ -0,0 +1,114 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "FGP_dTV_core.h" + +/* C-OMP implementation of FGP-dTV [1,2] denoising/regularization model (2D/3D case) + * which employs structural similarity of the level sets of two images/volumes, see [1,2] + * The current implementation updates image 1 while image 2 is being fixed. + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. Additional reference image/volume of the same dimensions as (1) [REQUIRED] + * 3. lambdaPar - regularization parameter [REQUIRED] + * 4. Number of iterations [OPTIONAL] + * 5. eplsilon: tolerance constant [OPTIONAL] + * 6. eta: smoothing constant to calculate gradient of the reference [OPTIONAL] * + * 7. TV-type: methodTV - 'iso' (0) or 'l1' (1) [OPTIONAL] + * 8. nonneg: 'nonnegativity (0 is OFF by default) [OPTIONAL] + * 9. print information: 0 (off) or 1 (on) [OPTIONAL] + * + * Output: + * [1] Filtered/regularized image/volume + * + * This function is based on the Matlab's codes and papers by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + * [2] M. J. Ehrhardt and M. M. Betcke, Multi-Contrast MRI Reconstruction with Structure-Guided Total Variation, SIAM Journal on Imaging Sciences 9(3), pp. 1084–1106 + */ + + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter, methTV, printswitch, nonneg; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + const mwSize *dim_array2; + float *Input, *InputRef, *Output=NULL, lambda, epsil, eta; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + dim_array2 = mxGetDimensions(prhs[1]); + + /*Handling Matlab input data*/ + if ((nrhs < 3) || (nrhs > 9)) mexErrMsgTxt("At least 3 parameters is required, all parameters are: Image(2D/3D), Reference(2D/3D), Regularization parameter, iterations number, tolerance, smoothing constant, penalty type ('iso' or 'l1'), nonnegativity switch, print switch"); + + Input = (float *) mxGetData(prhs[0]); /*noisy image (2D/3D) */ + InputRef = (float *) mxGetData(prhs[1]); /* reference image (2D/3D) */ + lambda = (float) mxGetScalar(prhs[2]); /* regularization parameter */ + iter = 300; /* default iterations number */ + epsil = 0.0001; /* default tolerance constant */ + eta = 0.01; /* default smoothing constant */ + methTV = 0; /* default isotropic TV penalty */ + nonneg = 0; /* default nonnegativity switch, off - 0 */ + printswitch = 0; /*default print is switched, off - 0 */ + + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if (mxGetClassID(prhs[1]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + if (number_of_dims == 2) { if ((dimX != dim_array2[0]) || (dimY != dim_array2[1])) mexErrMsgTxt("The input images have different dimensionalities");} + if (number_of_dims == 3) { if ((dimX != dim_array2[0]) || (dimY != dim_array2[1]) || (dimZ != dim_array2[2])) mexErrMsgTxt("The input volumes have different dimensionalities");} + + + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6) || (nrhs == 7) || (nrhs == 8) || (nrhs == 9)) iter = (int) mxGetScalar(prhs[3]); /* iterations number */ + if ((nrhs == 5) || (nrhs == 6) || (nrhs == 7) || (nrhs == 8) || (nrhs == 9)) epsil = (float) mxGetScalar(prhs[4]); /* tolerance constant */ + if ((nrhs == 6) || (nrhs == 7) || (nrhs == 8) || (nrhs == 9)) { + eta = (float) mxGetScalar(prhs[5]); /* smoothing constant for the gradient of InputRef */ + } + if ((nrhs == 7) || (nrhs == 8) || (nrhs == 9)) { + char *penalty_type; + penalty_type = mxArrayToString(prhs[6]); /* choosing TV penalty: 'iso' or 'l1', 'iso' is the default */ + if ((strcmp(penalty_type, "l1") != 0) && (strcmp(penalty_type, "iso") != 0)) mexErrMsgTxt("Choose TV type: 'iso' or 'l1',"); + if (strcmp(penalty_type, "l1") == 0) methTV = 1; /* enable 'l1' penalty */ + mxFree(penalty_type); + } + if ((nrhs == 8) || (nrhs == 9)) { + nonneg = (int) mxGetScalar(prhs[7]); + if ((nonneg != 0) && (nonneg != 1)) mexErrMsgTxt("Nonnegativity constraint can be enabled by choosing 1 or off - 0"); + } + if (nrhs == 9) { + printswitch = (int) mxGetScalar(prhs[8]); + if ((printswitch != 0) && (printswitch != 1)) mexErrMsgTxt("Print can be enabled by choosing 1 or off - 0"); + } + + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + /* running the function */ + dTV_FGP_CPU_main(Input, InputRef, Output, lambda, iter, epsil, eta, methTV, nonneg, printswitch, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_CPU/LLT_ROF.c b/src/Matlab/mex_compile/regularisers_CPU/LLT_ROF.c new file mode 100644 index 0000000..ab45446 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/LLT_ROF.c @@ -0,0 +1,82 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "LLT_ROF_core.h" + +/* C-OMP implementation of Lysaker, Lundervold and Tai (LLT) model [1] combined with Rudin-Osher-Fatemi [2] TV regularisation penalty. +* +* This penalty can deliver visually pleasant piecewise-smooth recovery if regularisation parameters are selected well. +* The rule of thumb for selection is to start with lambdaLLT = 0 (just the ROF-TV model) and then proceed to increase +* lambdaLLT starting with smaller values. +* +* Input Parameters: +* 1. U0 - original noise image/volume +* 2. lambdaROF - ROF-related regularisation parameter +* 3. lambdaLLT - LLT-related regularisation parameter +* 4. tau - time-marching step +* 5. iter - iterations number (for both models) +* +* Output: +* Filtered/regularised image +* +* References: +* [1] Lysaker, M., Lundervold, A. and Tai, X.C., 2003. Noise removal using fourth-order partial differential equation with applications to medical magnetic resonance images in space and time. IEEE Transactions on image processing, 12(12), pp.1579-1590. +* [2] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" +*/ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iterationsNumb; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + float *Input, *Output=NULL, lambdaROF, lambdaLLT, tau; + + dim_array = mxGetDimensions(prhs[0]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + if ((nrhs < 3) || (nrhs > 5)) mexErrMsgTxt("At least 3 parameters is required, all parameters are: Image(2D/3D), Regularisation parameter (ROF), Regularisation parameter (LTT), iterations number, time-marching parameter"); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + lambdaROF = (float) mxGetScalar(prhs[1]); /* ROF regularization parameter */ + lambdaLLT = (float) mxGetScalar(prhs[2]); /* ROF regularization parameter */ + iterationsNumb = 250; + tau = 0.0025; + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if ((nrhs == 4) || (nrhs == 5)) iterationsNumb = (int) mxGetScalar(prhs[3]); /* iterations number */ + if (nrhs == 5) tau = (float) mxGetScalar(prhs[4]); /* marching step parameter */ + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + LLT_ROF_CPU_main(Input, Output, lambdaROF, lambdaLLT, iterationsNumb, tau, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_CPU/NonlDiff.c b/src/Matlab/mex_compile/regularisers_CPU/NonlDiff.c new file mode 100644 index 0000000..ec35b8b --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/NonlDiff.c @@ -0,0 +1,89 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "Diffusion_core.h" + +/* C-OMP implementation of linear and nonlinear diffusion with the regularisation model [1] (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambda - regularization parameter + * 3. Edge-preserving parameter (sigma), when sigma equals to zero nonlinear diffusion -> linear diffusion + * 4. Number of iterations, for explicit scheme >= 150 is recommended [OPTIONAL parameter] + * 5. tau - time-marching step for explicit scheme [OPTIONAL parameter] + * 6. Penalty type: 1 - Huber, 2 - Perona-Malik, 3 - Tukey Biweight [OPTIONAL parameter] + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Perona, P. and Malik, J., 1990. Scale-space and edge detection using anisotropic diffusion. IEEE Transactions on pattern analysis and machine intelligence, 12(7), pp.629-639. + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter_numb, penaltytype; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + + float *Input, *Output=NULL, lambda, tau, sigma; + + dim_array = mxGetDimensions(prhs[0]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + sigma = (float) mxGetScalar(prhs[2]); /* Edge-preserving parameter */ + iter_numb = 300; /* iterations number */ + tau = 0.025; /* marching step parameter */ + penaltytype = 1; /* Huber penalty by default */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if ((nrhs < 3) || (nrhs > 6)) mexErrMsgTxt("At least 3 parameters is required, all parameters are: Image(2D/3D), Regularisation parameter, Edge-preserving parameter, iterations number, time-marching constant, penalty type - Huber, PM or Tukey"); + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6)) iter_numb = (int) mxGetScalar(prhs[3]); /* iterations number */ + if ((nrhs == 5) || (nrhs == 6)) tau = (float) mxGetScalar(prhs[4]); /* marching step parameter */ + if (nrhs == 6) { + char *penalty_type; + penalty_type = mxArrayToString(prhs[5]); /* Huber, PM or Tukey 'Huber' is the default */ + if ((strcmp(penalty_type, "Huber") != 0) && (strcmp(penalty_type, "PM") != 0) && (strcmp(penalty_type, "Tukey") != 0)) mexErrMsgTxt("Choose penalty: 'Huber', 'PM' or 'Tukey',"); + if (strcmp(penalty_type, "Huber") == 0) penaltytype = 1; /* enable 'Huber' penalty */ + if (strcmp(penalty_type, "PM") == 0) penaltytype = 2; /* enable Perona-Malik penalty */ + if (strcmp(penalty_type, "Tukey") == 0) penaltytype = 3; /* enable Tikey Biweight penalty */ + mxFree(penalty_type); + } + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + Diffusion_CPU_main(Input, Output, lambda, sigma, iter_numb, tau, penaltytype, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_CPU/NonlDiff_Inp.c b/src/Matlab/mex_compile/regularisers_CPU/NonlDiff_Inp.c new file mode 100644 index 0000000..9833392 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/NonlDiff_Inp.c @@ -0,0 +1,103 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "Diffusion_Inpaint_core.h" + +/* C-OMP implementation of linear and nonlinear diffusion [1,2] for inpainting task (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Image/volume to inpaint + * 2. Inpainting Mask of the same size as (1) in 'unsigned char' format (ones mark the region to inpaint, zeros belong to the data) + * 3. lambda - regularization parameter + * 4. Edge-preserving parameter (sigma), when sigma equals to zero nonlinear diffusion -> linear diffusion + * 5. Number of iterations, for explicit scheme >= 150 is recommended + * 6. tau - time-marching step for explicit scheme + * 7. Penalty type: 1 - Huber, 2 - Perona-Malik, 3 - Tukey Biweight + * + * Output: + * [1] Inpainted image/volume + * + * This function is based on the paper by + * [1] Perona, P. and Malik, J., 1990. Scale-space and edge detection using anisotropic diffusion. IEEE Transactions on pattern analysis and machine intelligence, 12(7), pp.629-639. + * [2] Black, M.J., Sapiro, G., Marimont, D.H. and Heeger, D., 1998. Robust anisotropic diffusion. IEEE Transactions on image processing, 7(3), pp.421-432. + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter_numb, penaltytype, i, inpaint_elements; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + const mwSize *dim_array2; + + float *Input, *Output=NULL, lambda, tau, sigma; + unsigned char *Mask; + + dim_array = mxGetDimensions(prhs[0]); + dim_array2 = mxGetDimensions(prhs[1]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + Mask = (unsigned char *) mxGetData(prhs[1]); /* MASK */ + lambda = (float) mxGetScalar(prhs[2]); /* regularization parameter */ + sigma = (float) mxGetScalar(prhs[3]); /* Edge-preserving parameter */ + iter_numb = 300; /* iterations number */ + tau = 0.025; /* marching step parameter */ + penaltytype = 1; /* Huber penalty by default */ + + if ((nrhs < 4) || (nrhs > 7)) mexErrMsgTxt("At least 4 parameters is required, all parameters are: Image(2D/3D), Mask(2D/3D), Regularisation parameter, Edge-preserving parameter, iterations number, time-marching constant, penalty type - Huber, PM or Tukey"); + if ((nrhs == 5) || (nrhs == 6) || (nrhs == 7)) iter_numb = (int) mxGetScalar(prhs[4]); /* iterations number */ + if ((nrhs == 6) || (nrhs == 7)) tau = (float) mxGetScalar(prhs[5]); /* marching step parameter */ + if (nrhs == 7) { + char *penalty_type; + penalty_type = mxArrayToString(prhs[6]); /* Huber, PM or Tukey 'Huber' is the default */ + if ((strcmp(penalty_type, "Huber") != 0) && (strcmp(penalty_type, "PM") != 0) && (strcmp(penalty_type, "Tukey") != 0)) mexErrMsgTxt("Choose penalty: 'Huber', 'PM' or 'Tukey',"); + if (strcmp(penalty_type, "Huber") == 0) penaltytype = 1; /* enable 'Huber' penalty */ + if (strcmp(penalty_type, "PM") == 0) penaltytype = 2; /* enable Perona-Malik penalty */ + if (strcmp(penalty_type, "Tukey") == 0) penaltytype = 3; /* enable Tikey Biweight penalty */ + mxFree(penalty_type); + } + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if (mxGetClassID(prhs[1]) != mxUINT8_CLASS) {mexErrMsgTxt("The mask must be in uint8 precision");} + + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + if ((dimX != dim_array2[0]) || (dimY != dim_array2[1])) mexErrMsgTxt("Input image and the provided mask are of different dimensions!"); + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) { + if ((dimX != dim_array2[0]) || (dimY != dim_array2[1]) || (dimZ != dim_array2[2])) mexErrMsgTxt("Input image and the provided mask are of different dimensions!"); + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + } + + inpaint_elements = 0; + for (i=0; i<(int)(dimY*dimX*dimZ); i++) if (Mask[i] == 1) inpaint_elements++; + if (inpaint_elements == 0) mexErrMsgTxt("The mask is full of zeros, nothing to inpaint"); + Diffusion_Inpaint_CPU_main(Input, Mask, Output, lambda, sigma, iter_numb, tau, penaltytype, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_CPU/NonlocalMarching_Inpaint.c b/src/Matlab/mex_compile/regularisers_CPU/NonlocalMarching_Inpaint.c new file mode 100644 index 0000000..b3f2c98 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/NonlocalMarching_Inpaint.c @@ -0,0 +1,84 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "NonlocalMarching_Inpaint_core.h" + +/* C-OMP implementation of Nonlocal Vertical Marching inpainting method (2D case) + * The method is heuristic but computationally efficent (especially for larger images). + * It developed specifically to smoothly inpaint horizontal or inclined missing data regions in sinograms + * The method WILL not work satisfactory if you have lengthy vertical stripes of missing data + * + * Input: + * 1. 2D image or sinogram [REQUIRED] + * 2. Mask of the same size as A in 'unsigned char' format (ones mark the region to inpaint, zeros belong to the data) [REQUIRED] + * 3. Linear increment to increase searching window size in iterations, values from 1-3 is a good choice [OPTIONAL, default 1] + * 4. Number of iterations [OPTIONAL, default - calculate based on the mask] + * + * Output: + * 1. Inpainted sinogram + * 2. updated mask + * Reference: TBA + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iterations, SW_increment; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + const mwSize *dim_array2; + + float *Input, *Output=NULL; + unsigned char *Mask, *Mask_upd=NULL; + + dim_array = mxGetDimensions(prhs[0]); + dim_array2 = mxGetDimensions(prhs[1]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + Mask = (unsigned char *) mxGetData(prhs[1]); /* MASK */ + SW_increment = 1; + iterations = 0; + + if ((nrhs < 2) || (nrhs > 4)) mexErrMsgTxt("At least 4 parameters is required, all parameters are: Image(2D/3D), Mask(2D/3D), Linear increment, Iterations number"); + if ((nrhs == 3) || (nrhs == 4)) SW_increment = (int) mxGetScalar(prhs[2]); /* linear increment */ + if ((nrhs == 4)) iterations = (int) mxGetScalar(prhs[3]); /* iterations number */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if (mxGetClassID(prhs[1]) != mxUINT8_CLASS) {mexErrMsgTxt("The mask must be in uint8 precision");} + + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + if ((dimX != dim_array2[0]) || (dimY != dim_array2[1])) mexErrMsgTxt("Input image and the provided mask are of different dimensions!"); + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + Mask_upd = (unsigned char*)mxGetPr(plhs[1] = mxCreateNumericArray(2, dim_array, mxUINT8_CLASS, mxREAL)); + } + if (number_of_dims == 3) { + mexErrMsgTxt("Currently 2D supported only"); + } + NonlocalMarching_Inpaint_main(Input, Mask, Output, Mask_upd, SW_increment, iterations, 0, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV.c b/src/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV.c new file mode 100644 index 0000000..014c0a0 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV.c @@ -0,0 +1,88 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC and Diamond Light Source Ltd. + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * Copyright 2018 Diamond Light Source Ltd. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "matrix.h" +#include "mex.h" +#include "Nonlocal_TV_core.h" + +#define EPS 1.0000e-9 + +/* Matlab wrapper for C-OMP implementation of non-local regulariser + * Weights and associated indices must be given as an input. + * Gauss-Seidel fixed point iteration requires ~ 3 iterations, so the main effort + * goes in pre-calculation of weights and selection of patches + * + * + * Input Parameters: + * 1. 2D/3D grayscale image/volume + * 2. AR_i - indeces of i neighbours + * 3. AR_j - indeces of j neighbours + * 4. AR_k - indeces of k neighbours (0 - for 2D case) + * 5. Weights_ij(k) - associated weights + * 6. regularisation parameter + * 7. iterations number + + * Output: + * 1. denoised image/volume + * Elmoataz, Abderrahim, Olivier Lezoray, and Sébastien Bougleux. "Nonlocal discrete regularization on weighted graphs: a framework for image and manifold processing." IEEE Trans. Image Processing 17, no. 7 (2008): 1047-1060. + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) +{ + long number_of_dims, dimX, dimY, dimZ; + int IterNumb, NumNeighb = 0; + unsigned short *H_i, *H_j, *H_k; + const int *dim_array; + const int *dim_array2; + float *A_orig, *Output=NULL, *Weights, lambda; + + dim_array = mxGetDimensions(prhs[0]); + dim_array2 = mxGetDimensions(prhs[1]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + A_orig = (float *) mxGetData(prhs[0]); /* a 2D image or a set of 2D images (3D stack) */ + H_i = (unsigned short *) mxGetData(prhs[1]); /* indeces of i neighbours */ + H_j = (unsigned short *) mxGetData(prhs[2]); /* indeces of j neighbours */ + H_k = (unsigned short *) mxGetData(prhs[3]); /* indeces of k neighbours */ + Weights = (float *) mxGetData(prhs[4]); /* weights for patches */ + lambda = (float) mxGetScalar(prhs[5]); /* regularisation parameter */ + IterNumb = (int) mxGetScalar(prhs[6]); /* the number of iterations */ + + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /*****2D INPUT *****/ + if (number_of_dims == 2) { + dimZ = 0; + NumNeighb = dim_array2[2]; + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + /*****3D INPUT *****/ + /****************************************************/ + if (number_of_dims == 3) { + NumNeighb = dim_array2[3]; + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + } + + /* run the main function here */ + Nonlocal_TV_CPU_main(A_orig, Output, H_i, H_j, H_k, Weights, dimX, dimY, dimZ, NumNeighb, lambda, IterNumb); +} diff --git a/src/Matlab/mex_compile/regularisers_CPU/PatchSelect.c b/src/Matlab/mex_compile/regularisers_CPU/PatchSelect.c new file mode 100644 index 0000000..f942539 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/PatchSelect.c @@ -0,0 +1,92 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC and Diamond Light Source Ltd. + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * Copyright 2018 Diamond Light Source Ltd. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#include "matrix.h" +#include "mex.h" +#include "PatchSelect_core.h" + +/* C-OMP implementation of non-local weight pre-calculation for non-local priors + * Weights and associated indices are stored into pre-allocated arrays and passed + * to the regulariser + * + * + * Input Parameters: + * 1. 2D/3D grayscale image/volume + * 2. Searching window (half-size of the main bigger searching window, e.g. 11) + * 3. Similarity window (half-size of the patch window, e.g. 2) + * 4. The number of neighbours to take (the most prominent after sorting neighbours will be taken) + * 5. noise-related parameter to calculate non-local weights + * + * Output [2D]: + * 1. AR_i - indeces of i neighbours + * 2. AR_j - indeces of j neighbours + * 3. Weights_ij - associated weights + * + * Output [3D]: + * 1. AR_i - indeces of i neighbours + * 2. AR_j - indeces of j neighbours + * 3. AR_k - indeces of j neighbours + * 4. Weights_ijk - associated weights + */ +/**************************************************/ +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) +{ + int number_of_dims, SearchWindow, SimilarWin, NumNeighb; + mwSize dimX, dimY, dimZ; + unsigned short *H_i=NULL, *H_j=NULL, *H_k=NULL; + const int *dim_array; + float *A, *Weights = NULL, h; + int dim_array2[3]; /* for 2D data */ + int dim_array3[4]; /* for 3D data */ + + dim_array = mxGetDimensions(prhs[0]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + A = (float *) mxGetData(prhs[0]); /* a 2D or 3D image/volume */ + SearchWindow = (int) mxGetScalar(prhs[1]); /* Large Searching window */ + SimilarWin = (int) mxGetScalar(prhs[2]); /* Similarity window (patch-search)*/ + NumNeighb = (int) mxGetScalar(prhs[3]); /* the total number of neighbours to take */ + h = (float) mxGetScalar(prhs[4]); /* NLM parameter */ + + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + dim_array2[0] = dimX; dim_array2[1] = dimY; dim_array2[2] = NumNeighb; /* 2D case */ + dim_array3[0] = dimX; dim_array3[1] = dimY; dim_array3[2] = dimZ; dim_array3[3] = NumNeighb; /* 3D case */ + + /****************2D INPUT ***************/ + if (number_of_dims == 2) { + dimZ = 0; + H_i = (unsigned short*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array2, mxUINT16_CLASS, mxREAL)); + H_j = (unsigned short*)mxGetPr(plhs[1] = mxCreateNumericArray(3, dim_array2, mxUINT16_CLASS, mxREAL)); + Weights = (float*)mxGetPr(plhs[2] = mxCreateNumericArray(3, dim_array2, mxSINGLE_CLASS, mxREAL)); + } + /****************3D INPUT ***************/ + if (number_of_dims == 3) { + H_i = (unsigned short*)mxGetPr(plhs[0] = mxCreateNumericArray(4, dim_array3, mxUINT16_CLASS, mxREAL)); + H_j = (unsigned short*)mxGetPr(plhs[1] = mxCreateNumericArray(4, dim_array3, mxUINT16_CLASS, mxREAL)); + H_k = (unsigned short*)mxGetPr(plhs[2] = mxCreateNumericArray(4, dim_array3, mxUINT16_CLASS, mxREAL)); + Weights = (float*)mxGetPr(plhs[3] = mxCreateNumericArray(4, dim_array3, mxSINGLE_CLASS, mxREAL)); + } + + PatchSelect_CPU_main(A, H_i, H_j, H_k, Weights, (long)(dimX), (long)(dimY), (long)(dimZ), SearchWindow, SimilarWin, NumNeighb, h, 0); + + } diff --git a/src/Matlab/mex_compile/regularisers_CPU/ROF_TV.c b/src/Matlab/mex_compile/regularisers_CPU/ROF_TV.c new file mode 100644 index 0000000..55ef2b1 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/ROF_TV.c @@ -0,0 +1,77 @@ + +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "ROF_TV_core.h" + +/* ROF-TV denoising/regularization model [1] (2D/3D case) + * (MEX wrapper for MATLAB) + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. lambda - regularization parameter [REQUIRED] + * 3. Number of iterations, for explicit scheme >= 150 is recommended [REQUIRED] + * 4. tau - marching step for explicit scheme, ~1 is recommended [REQUIRED] + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" + * + * D. Kazantsev, 2016-18 + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter_numb; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array_i; + float *Input, *Output=NULL, lambda, tau; + + dim_array_i = mxGetDimensions(prhs[0]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + iter_numb = (int) mxGetScalar(prhs[2]); /* iterations number */ + tau = (float) mxGetScalar(prhs[3]); /* marching step parameter */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if(nrhs != 4) mexErrMsgTxt("Four inputs reqired: Image(2D,3D), regularization parameter, iterations number, marching step constant"); + /*Handling Matlab output data*/ + dimX = dim_array_i[0]; dimY = dim_array_i[1]; dimZ = dim_array_i[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array_i, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) { + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array_i, mxSINGLE_CLASS, mxREAL)); + } + + TV_ROF_CPU_main(Input, Output, lambda, iter_numb, tau, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_CPU/SB_TV.c b/src/Matlab/mex_compile/regularisers_CPU/SB_TV.c new file mode 100644 index 0000000..8636322 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/SB_TV.c @@ -0,0 +1,91 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "SB_TV_core.h" + +/* C-OMP implementation of Split Bregman - TV denoising-regularisation model (2D/3D) [1] +* +* Input Parameters: +* 1. Noisy image/volume +* 2. lambda - regularisation parameter +* 3. Number of iterations [OPTIONAL parameter] +* 4. eplsilon - tolerance constant [OPTIONAL parameter] +* 5. TV-type: 'iso' or 'l1' [OPTIONAL parameter] +* 6. print information: 0 (off) or 1 (on) [OPTIONAL parameter] +* +* Output: +* 1. Filtered/regularized image +* +* This function is based on the Matlab's code and paper by +* [1]. Goldstein, T. and Osher, S., 2009. The split Bregman method for L1-regularized problems. SIAM journal on imaging sciences, 2(2), pp.323-343. +*/ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter, methTV, printswitch; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + + float *Input, *Output=NULL, lambda, epsil; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + + /*Handling Matlab input data*/ + if ((nrhs < 2) || (nrhs > 6)) mexErrMsgTxt("At least 2 parameters is required, all parameters are: Image(2D/3D), Regularization parameter, Regularization parameter, iterations number, tolerance, penalty type ('iso' or 'l1'), print switch"); + + Input = (float *) mxGetData(prhs[0]); /*noisy image (2D/3D) */ + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + iter = 100; /* default iterations number */ + epsil = 0.0001; /* default tolerance constant */ + methTV = 0; /* default isotropic TV penalty */ + printswitch = 0; /*default print is switched, off - 0 */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + + if ((nrhs == 3) || (nrhs == 4) || (nrhs == 5) || (nrhs == 6)) iter = (int) mxGetScalar(prhs[2]); /* iterations number */ + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6)) epsil = (float) mxGetScalar(prhs[3]); /* tolerance constant */ + if ((nrhs == 5) || (nrhs == 6)) { + char *penalty_type; + penalty_type = mxArrayToString(prhs[4]); /* choosing TV penalty: 'iso' or 'l1', 'iso' is the default */ + if ((strcmp(penalty_type, "l1") != 0) && (strcmp(penalty_type, "iso") != 0)) mexErrMsgTxt("Choose TV type: 'iso' or 'l1',"); + if (strcmp(penalty_type, "l1") == 0) methTV = 1; /* enable 'l1' penalty */ + mxFree(penalty_type); + } + if (nrhs == 6) { + printswitch = (int) mxGetScalar(prhs[5]); + if ((printswitch != 0) && (printswitch != 1)) mexErrMsgTxt("Print can be enabled by choosing 1 or off - 0"); + } + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + /* running the function */ + SB_TV_CPU_main(Input, Output, lambda, iter, epsil, methTV, printswitch, dimX, dimY, dimZ); +} diff --git a/src/Matlab/mex_compile/regularisers_CPU/TGV.c b/src/Matlab/mex_compile/regularisers_CPU/TGV.c new file mode 100644 index 0000000..aa4eed4 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/TGV.c @@ -0,0 +1,83 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "mex.h" +#include "TGV_core.h" + +/* C-OMP implementation of Primal-Dual denoising method for + * Total Generilized Variation (TGV)-L2 model [1] (2D/3D) + * + * Input Parameters: + * 1. Noisy image/volume (2D/3D) + * 2. lambda - regularisation parameter + * 3. parameter to control the first-order term (alpha1) + * 4. parameter to control the second-order term (alpha0) + * 5. Number of Chambolle-Pock (Primal-Dual) iterations + * 6. Lipshitz constant (default is 12) + * + * Output: + * Filtered/regulariaed image + * + * References: + * [1] K. Bredies "Total Generalized Variation" + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + + float *Input, *Output=NULL, lambda, alpha0, alpha1, L2; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + + /*Handling Matlab input data*/ + if ((nrhs < 2) || (nrhs > 6)) mexErrMsgTxt("At least 2 parameters is required, all parameters are: Image(2D), Regularisation parameter, alpha0, alpha1, iterations number, Lipshitz Constant"); + + Input = (float *) mxGetData(prhs[0]); /*noisy image/volume */ + lambda = (float) mxGetScalar(prhs[1]); /* regularisation parameter */ + alpha1 = 1.0f; /* parameter to control the first-order term */ + alpha0 = 0.5f; /* parameter to control the second-order term */ + iter = 300; /* Iterations number */ + L2 = 12.0f; /* Lipshitz constant */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if ((nrhs == 3) || (nrhs == 4) || (nrhs == 5) || (nrhs == 6)) alpha1 = (float) mxGetScalar(prhs[2]); /* parameter to control the first-order term */ + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6)) alpha0 = (float) mxGetScalar(prhs[3]); /* parameter to control the second-order term */ + if ((nrhs == 5) || (nrhs == 6)) iter = (int) mxGetScalar(prhs[4]); /* Iterations number */ + if (nrhs == 6) L2 = (float) mxGetScalar(prhs[5]); /* Lipshitz constant */ + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) { + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + } + /* running the function */ + TGV_main(Input, Output, lambda, alpha1, alpha0, iter, L2, dimX, dimY, dimZ); +} diff --git a/src/Matlab/mex_compile/regularisers_CPU/TNV.c b/src/Matlab/mex_compile/regularisers_CPU/TNV.c new file mode 100644 index 0000000..acea75d --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/TNV.c @@ -0,0 +1,74 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "TNV_core.h" +/* + * C-OMP implementation of Total Nuclear Variation regularisation model (2D + channels) [1] + * The code is modified from the implementation by Joan Duran <joan.duran@uib.es> see + * "denoisingPDHG_ipol.cpp" in Joans Collaborative Total Variation package + * + * Input Parameters: + * 1. Noisy volume of 2D + channel dimension, i.e. 3D volume + * 2. lambda - regularisation parameter + * 3. Number of iterations [OPTIONAL parameter] + * 4. eplsilon - tolerance constant [OPTIONAL parameter] + * 5. print information: 0 (off) or 1 (on) [OPTIONAL parameter] + * + * Output: + * 1. Filtered/regularized image + * + * [1]. Duran, J., Moeller, M., Sbert, C. and Cremers, D., 2016. Collaborative total variation: a general framework for vectorial TV models. SIAM Journal on Imaging Sciences, 9(1), pp.116-151. + */ +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + float *Input, *Output=NULL, lambda, epsil; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + + /*Handling Matlab input data*/ + if ((nrhs < 2) || (nrhs > 4)) mexErrMsgTxt("At least 2 parameters is required, all parameters are: Image(2D + channels), Regularisation parameter, Regularization parameter, iterations number, tolerance"); + + Input = (float *) mxGetData(prhs[0]); /* noisy sequence of channels (2D + channels) */ + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + iter = 1000; /* default iterations number */ + epsil = 1.00e-05; /* default tolerance constant */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + + if ((nrhs == 3) || (nrhs == 4)) iter = (int) mxGetScalar(prhs[2]); /* iterations number */ + if (nrhs == 4) epsil = (float) mxGetScalar(prhs[3]); /* tolerance constant */ + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + if (number_of_dims == 2) mexErrMsgTxt("The input must be 3D: [X,Y,Channels]"); + if (number_of_dims == 3) { + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + /* running the function */ + TNV_CPU_main(Input, Output, lambda, iter, epsil, dimX, dimY, dimZ); + } +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_CPU/TV_energy.c b/src/Matlab/mex_compile/regularisers_CPU/TV_energy.c new file mode 100644 index 0000000..d457f46 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_CPU/TV_energy.c @@ -0,0 +1,72 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "utils.h" +/* + * Function to calculate TV energy value with respect to the denoising variational problem + * + * Input: + * 1. Denoised Image/volume + * 2. Original (noisy) Image/volume + * 3. lambda - regularisation parameter + * + * Output: + * 1. Energy function value + * + */ +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, type; + + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + float *Input, *Input0, lambda; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + + /*Handling Matlab input data*/ + if ((nrhs != 4)) mexErrMsgTxt("4 inputs: Two images or volumes of the same size required, estimated and the original (noisy), regularisation parameter, type"); + + Input = (float *) mxGetData(prhs[0]); /* Denoised Image/volume */ + Input0 = (float *) mxGetData(prhs[1]); /* Original (noisy) Image/volume */ + lambda = (float) mxGetScalar(prhs[2]); /* regularisation parameter */ + type = (int) mxGetScalar(prhs[3]); /* type of energy */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if (mxGetClassID(prhs[1]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + + /*output energy function value */ + plhs[0] = mxCreateNumericMatrix(1, 1, mxSINGLE_CLASS, mxREAL); + float *funcvalA = (float *) mxGetData(plhs[0]); + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + if (number_of_dims == 2) { + TV_energy2D(Input, Input0, funcvalA, lambda, type, dimX, dimY); + } + if (number_of_dims == 3) { + TV_energy3D(Input, Input0, funcvalA, lambda, type, dimX, dimY, dimZ); + } +} diff --git a/src/Matlab/mex_compile/regularisers_GPU/Diffusion_4thO_GPU.cpp b/src/Matlab/mex_compile/regularisers_GPU/Diffusion_4thO_GPU.cpp new file mode 100644 index 0000000..0cc042b --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_GPU/Diffusion_4thO_GPU.cpp @@ -0,0 +1,77 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "Diffus_4thO_GPU_core.h" + +/* CUDA implementation of fourth-order diffusion scheme [1] for piecewise-smooth recovery (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. lambda - regularization parameter [REQUIRED] + * 3. Edge-preserving parameter (sigma) [REQUIRED] + * 4. Number of iterations, for explicit scheme >= 150 is recommended [OPTIONAL, default 300] + * 5. tau - time-marching step for the explicit scheme [OPTIONAL, default 0.015] + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Hajiaboli, M.R., 2011. An anisotropic fourth-order diffusion filter for image noise removal. International Journal of Computer Vision, 92(2), pp.177-191. + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter_numb; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + float *Input, *Output=NULL, lambda, tau, sigma; + + dim_array = mxGetDimensions(prhs[0]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + sigma = (float) mxGetScalar(prhs[2]); /* Edge-preserving parameter */ + iter_numb = 300; /* iterations number */ + tau = 0.01; /* marching step parameter */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if ((nrhs < 3) || (nrhs > 5)) mexErrMsgTxt("At least 3 parameters is required, all parameters are: Image(2D/3D), Regularisation parameter, Edge-preserving parameter, iterations number, time-marching constant"); + if ((nrhs == 4) || (nrhs == 5)) iter_numb = (int) mxGetScalar(prhs[3]); /* iterations number */ + if (nrhs == 5) tau = (float) mxGetScalar(prhs[4]); /* marching step parameter */ + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + Diffus4th_GPU_main(Input, Output, lambda, sigma, iter_numb, tau, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_GPU/FGP_TV_GPU.cpp b/src/Matlab/mex_compile/regularisers_GPU/FGP_TV_GPU.cpp new file mode 100644 index 0000000..c174e75 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_GPU/FGP_TV_GPU.cpp @@ -0,0 +1,97 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "TV_FGP_GPU_core.h" + +/* GPU (CUDA) implementation of FGP-TV [1] denoising/regularization model (2D/3D case) + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambdaPar - regularization parameter + * 3. Number of iterations + * 4. eplsilon: tolerance constant + * 5. TV-type: methodTV - 'iso' (0) or 'l1' (1) + * 6. nonneg: 'nonnegativity (0 is OFF by default) + * 7. print information: 0 (off) or 1 (on) + * + * Output: + * [1] Filtered/regularized image + * + * This function is based on the Matlab's code and paper by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter, methTV, printswitch, nonneg; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + + float *Input, *Output=NULL, lambda, epsil; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + + /*Handling Matlab input data*/ + if ((nrhs < 2) || (nrhs > 7)) mexErrMsgTxt("At least 2 parameters is required, all parameters are: Image(2D/3D), Regularization parameter. The full list of parameters: Image(2D/3D), Regularization parameter, iterations number, tolerance, penalty type ('iso' or 'l1'), nonnegativity switch, print switch"); + + Input = (float *) mxGetData(prhs[0]); /*noisy image (2D/3D) */ + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + iter = 300; /* default iterations number */ + epsil = 0.0001; /* default tolerance constant */ + methTV = 0; /* default isotropic TV penalty */ + nonneg = 0; /* default nonnegativity switch, off - 0 */ + printswitch = 0; /*default print is switched, off - 0 */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + + if ((nrhs == 3) || (nrhs == 4) || (nrhs == 5) || (nrhs == 6) || (nrhs == 7)) iter = (int) mxGetScalar(prhs[2]); /* iterations number */ + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6) || (nrhs == 7)) epsil = (float) mxGetScalar(prhs[3]); /* tolerance constant */ + if ((nrhs == 5) || (nrhs == 6) || (nrhs == 7)) { + char *penalty_type; + penalty_type = mxArrayToString(prhs[4]); /* choosing TV penalty: 'iso' or 'l1', 'iso' is the default */ + if ((strcmp(penalty_type, "l1") != 0) && (strcmp(penalty_type, "iso") != 0)) mexErrMsgTxt("Choose TV type: 'iso' or 'l1',"); + if (strcmp(penalty_type, "l1") == 0) methTV = 1; /* enable 'l1' penalty */ + mxFree(penalty_type); + } + if ((nrhs == 6) || (nrhs == 7)) { + nonneg = (int) mxGetScalar(prhs[5]); + if ((nonneg != 0) && (nonneg != 1)) mexErrMsgTxt("Nonnegativity constraint can be enabled by choosing 1 or off - 0"); + } + if (nrhs == 7) { + printswitch = (int) mxGetScalar(prhs[6]); + if ((printswitch != 0) && (printswitch != 1)) mexErrMsgTxt("Print can be enabled by choosing 1 or off - 0"); + } + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + /* running the function */ + TV_FGP_GPU_main(Input, Output, lambda, iter, epsil, methTV, nonneg, printswitch, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_GPU/FGP_dTV_GPU.cpp b/src/Matlab/mex_compile/regularisers_GPU/FGP_dTV_GPU.cpp new file mode 100644 index 0000000..3f5a4b3 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_GPU/FGP_dTV_GPU.cpp @@ -0,0 +1,113 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "dTV_FGP_GPU_core.h" + +/* CUDA implementation of FGP-dTV [1,2] denoising/regularization model (2D/3D case) + * which employs structural similarity of the level sets of two images/volumes, see [1,2] + * The current implementation updates image 1 while image 2 is being fixed. + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. Additional reference image/volume of the same dimensions as (1) [REQUIRED] + * 3. lambdaPar - regularization parameter [REQUIRED] + * 4. Number of iterations [OPTIONAL] + * 5. eplsilon: tolerance constant [OPTIONAL] + * 6. eta: smoothing constant to calculate gradient of the reference [OPTIONAL] * + * 7. TV-type: methodTV - 'iso' (0) or 'l1' (1) [OPTIONAL] + * 8. nonneg: 'nonnegativity (0 is OFF by default) [OPTIONAL] + * 9. print information: 0 (off) or 1 (on) [OPTIONAL] + * + * Output: + * [1] Filtered/regularized image/volume + * + * This function is based on the Matlab's codes and papers by + * [1] Amir Beck and Marc Teboulle, "Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems" + * [2] M. J. Ehrhardt and M. M. Betcke, Multi-Contrast MRI Reconstruction with Structure-Guided Total Variation, SIAM Journal on Imaging Sciences 9(3), pp. 1084–1106 + */ +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter, methTV, printswitch, nonneg; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + const mwSize *dim_array2; + + float *Input, *InputRef, *Output=NULL, lambda, epsil, eta; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + dim_array2 = mxGetDimensions(prhs[1]); + + /*Handling Matlab input data*/ + if ((nrhs < 3) || (nrhs > 9)) mexErrMsgTxt("At least 3 parameters is required, all parameters are: Image(2D/3D), Reference(2D/3D), Regularization parameter, iterations number, tolerance, smoothing constant, penalty type ('iso' or 'l1'), nonnegativity switch, print switch"); + + Input = (float *) mxGetData(prhs[0]); /*noisy image (2D/3D) */ + InputRef = (float *) mxGetData(prhs[1]); /* reference image (2D/3D) */ + lambda = (float) mxGetScalar(prhs[2]); /* regularization parameter */ + iter = 300; /* default iterations number */ + epsil = 0.0001; /* default tolerance constant */ + eta = 0.01; /* default smoothing constant */ + methTV = 0; /* default isotropic TV penalty */ + nonneg = 0; /* default nonnegativity switch, off - 0 */ + printswitch = 0; /*default print is switched, off - 0 */ + + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if (mxGetClassID(prhs[1]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + if (number_of_dims == 2) { if ((dimX != dim_array2[0]) || (dimY != dim_array2[1])) mexErrMsgTxt("The input images have different dimensionalities");} + if (number_of_dims == 3) { if ((dimX != dim_array2[0]) || (dimY != dim_array2[1]) || (dimZ != dim_array2[2])) mexErrMsgTxt("The input volumes have different dimensionalities");} + + + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6) || (nrhs == 7) || (nrhs == 8) || (nrhs == 9)) iter = (int) mxGetScalar(prhs[3]); /* iterations number */ + if ((nrhs == 5) || (nrhs == 6) || (nrhs == 7) || (nrhs == 8) || (nrhs == 9)) epsil = (float) mxGetScalar(prhs[4]); /* tolerance constant */ + if ((nrhs == 6) || (nrhs == 7) || (nrhs == 8) || (nrhs == 9)) { + eta = (float) mxGetScalar(prhs[5]); /* smoothing constant for the gradient of InputRef */ + } + if ((nrhs == 7) || (nrhs == 8) || (nrhs == 9)) { + char *penalty_type; + penalty_type = mxArrayToString(prhs[6]); /* choosing TV penalty: 'iso' or 'l1', 'iso' is the default */ + if ((strcmp(penalty_type, "l1") != 0) && (strcmp(penalty_type, "iso") != 0)) mexErrMsgTxt("Choose TV type: 'iso' or 'l1',"); + if (strcmp(penalty_type, "l1") == 0) methTV = 1; /* enable 'l1' penalty */ + mxFree(penalty_type); + } + if ((nrhs == 8) || (nrhs == 9)) { + nonneg = (int) mxGetScalar(prhs[7]); + if ((nonneg != 0) && (nonneg != 1)) mexErrMsgTxt("Nonnegativity constraint can be enabled by choosing 1 or off - 0"); + } + if (nrhs == 9) { + printswitch = (int) mxGetScalar(prhs[8]); + if ((printswitch != 0) && (printswitch != 1)) mexErrMsgTxt("Print can be enabled by choosing 1 or off - 0"); + } + + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + /* running the function */ + dTV_FGP_GPU_main(Input, InputRef, Output, lambda, iter, epsil, eta, methTV, nonneg, printswitch, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_GPU/LLT_ROF_GPU.cpp b/src/Matlab/mex_compile/regularisers_GPU/LLT_ROF_GPU.cpp new file mode 100644 index 0000000..e8da4ce --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_GPU/LLT_ROF_GPU.cpp @@ -0,0 +1,83 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "LLT_ROF_GPU_core.h" + +/* CUDA implementation of Lysaker, Lundervold and Tai (LLT) model [1] combined with Rudin-Osher-Fatemi [2] TV regularisation penalty. +* +* This penalty can deliver visually pleasant piecewise-smooth recovery if regularisation parameters are selected well. +* The rule of thumb for selection is to start with lambdaLLT = 0 (just the ROF-TV model) and then proceed to increase +* lambdaLLT starting with smaller values. +* +* Input Parameters: +* 1. U0 - original noise image/volume +* 2. lambdaROF - ROF-related regularisation parameter +* 3. lambdaLLT - LLT-related regularisation parameter +* 4. tau - time-marching step +* 5. iter - iterations number (for both models) +* +* Output: +* Filtered/regularised image +* +* References: +* [1] Lysaker, M., Lundervold, A. and Tai, X.C., 2003. Noise removal using fourth-order partial differential equation with applications to medical magnetic resonance images in space and time. IEEE Transactions on image processing, 12(12), pp.1579-1590. +* [2] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" +*/ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iterationsNumb; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + + float *Input, *Output=NULL, lambdaROF, lambdaLLT, tau; + + dim_array = mxGetDimensions(prhs[0]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + if ((nrhs < 3) || (nrhs > 5)) mexErrMsgTxt("At least 3 parameters is required, all parameters are: Image(2D/3D), Regularisation parameter (ROF), Regularisation parameter (LTT), iterations number, time-marching parameter"); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + lambdaROF = (float) mxGetScalar(prhs[1]); /* ROF regularization parameter */ + lambdaLLT = (float) mxGetScalar(prhs[2]); /* ROF regularization parameter */ + iterationsNumb = 250; + tau = 0.0025; + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if ((nrhs == 4) || (nrhs == 5)) iterationsNumb = (int) mxGetScalar(prhs[3]); /* iterations number */ + if (nrhs == 5) tau = (float) mxGetScalar(prhs[4]); /* marching step parameter */ + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + LLT_ROF_GPU_main(Input, Output, lambdaROF, lambdaLLT, iterationsNumb, tau, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_GPU/NonlDiff_GPU.cpp b/src/Matlab/mex_compile/regularisers_GPU/NonlDiff_GPU.cpp new file mode 100644 index 0000000..1cd0cdc --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_GPU/NonlDiff_GPU.cpp @@ -0,0 +1,92 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include <stdio.h> +#include <string.h> +#include "NonlDiff_GPU_core.h" + +/* CUDA implementation of linear and nonlinear diffusion with the regularisation model [1,2] (2D/3D case) + * The minimisation is performed using explicit scheme. + * + * Input Parameters: + * 1. Noisy image/volume + * 2. lambda - regularization parameter + * 3. Edge-preserving parameter (sigma), when sigma equals to zero nonlinear diffusion -> linear diffusion + * 4. Number of iterations, for explicit scheme >= 150 is recommended + * 5. tau - time-marching step for explicit scheme + * 6. Penalty type: 1 - Huber, 2 - Perona-Malik, 3 - Tukey Biweight + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Perona, P. and Malik, J., 1990. Scale-space and edge detection using anisotropic diffusion. IEEE Transactions on pattern analysis and machine intelligence, 12(7), pp.629-639. + * [2] Black, M.J., Sapiro, G., Marimont, D.H. and Heeger, D., 1998. Robust anisotropic diffusion. IEEE Transactions on image processing, 7(3), pp.421-432. + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter_numb, penaltytype; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + + float *Input, *Output=NULL, lambda, tau, sigma; + + dim_array = mxGetDimensions(prhs[0]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + sigma = (float) mxGetScalar(prhs[2]); /* Edge-preserving parameter */ + iter_numb = 300; /* iterations number */ + tau = 0.025; /* marching step parameter */ + penaltytype = 1; /* Huber penalty by default */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if ((nrhs < 3) || (nrhs > 6)) mexErrMsgTxt("At least 3 parameters is required, all parameters are: Image(2D/3D), Regularisation parameter, Edge-preserving parameter, iterations number, time-marching constant, penalty type - Huber, PM or Tukey"); + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6)) iter_numb = (int) mxGetScalar(prhs[3]); /* iterations number */ + if ((nrhs == 5) || (nrhs == 6)) tau = (float) mxGetScalar(prhs[4]); /* marching step parameter */ + if (nrhs == 6) { + char *penalty_type; + penalty_type = mxArrayToString(prhs[5]); /* Huber, PM or Tukey 'Huber' is the default */ + if ((strcmp(penalty_type, "Huber") != 0) && (strcmp(penalty_type, "PM") != 0) && (strcmp(penalty_type, "Tukey") != 0)) mexErrMsgTxt("Choose penalty: 'Huber', 'PM' or 'Tukey',"); + if (strcmp(penalty_type, "Huber") == 0) penaltytype = 1; /* enable 'Huber' penalty */ + if (strcmp(penalty_type, "PM") == 0) penaltytype = 2; /* enable Perona-Malik penalty */ + if (strcmp(penalty_type, "Tukey") == 0) penaltytype = 3; /* enable Tikey Biweight penalty */ + mxFree(penalty_type); + } + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + NonlDiff_GPU_main(Input, Output, lambda, sigma, iter_numb, tau, penaltytype, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_GPU/ROF_TV_GPU.cpp b/src/Matlab/mex_compile/regularisers_GPU/ROF_TV_GPU.cpp new file mode 100644 index 0000000..bd01d55 --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_GPU/ROF_TV_GPU.cpp @@ -0,0 +1,74 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "TV_ROF_GPU_core.h" + +/* ROF-TV denoising/regularization model [1] (2D/3D case) + * (MEX wrapper for MATLAB) + * + * Input Parameters: + * 1. Noisy image/volume [REQUIRED] + * 2. lambda - regularization parameter [REQUIRED] + * 3. Number of iterations, for explicit scheme >= 150 is recommended [REQUIRED] + * 4. tau - marching step for explicit scheme, ~1 is recommended [REQUIRED] + * + * Output: + * [1] Regularized image/volume + * + * This function is based on the paper by + * [1] Rudin, Osher, Fatemi, "Nonlinear Total Variation based noise removal algorithms" + * + * D. Kazantsev, 2016-18 + */ +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter_numb; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + + float *Input, *Output=NULL, lambda, tau; + + dim_array = mxGetDimensions(prhs[0]); + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + + /*Handling Matlab input data*/ + Input = (float *) mxGetData(prhs[0]); + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + iter_numb = (int) mxGetScalar(prhs[2]); /* iterations number */ + tau = (float) mxGetScalar(prhs[3]); /* marching step parameter */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if(nrhs != 4) mexErrMsgTxt("Four inputs reqired: Image(2D,3D), regularization parameter, iterations number, marching step constant"); + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + /* output arrays*/ + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + /* output image/volume */ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + TV_ROF_GPU_main(Input, Output, lambda, iter_numb, tau, dimX, dimY, dimZ); +}
\ No newline at end of file diff --git a/src/Matlab/mex_compile/regularisers_GPU/SB_TV_GPU.cpp b/src/Matlab/mex_compile/regularisers_GPU/SB_TV_GPU.cpp new file mode 100644 index 0000000..9d1328f --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_GPU/SB_TV_GPU.cpp @@ -0,0 +1,91 @@ +/* + * This work is part of the Core Imaging Library developed by + * Visual Analytics and Imaging System Group of the Science Technology + * Facilities Council, STFC + * + * Copyright 2017 Daniil Kazantsev + * Copyright 2017 Srikanth Nagella, Edoardo Pasca + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * http://www.apache.org/licenses/LICENSE-2.0 + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +#include "matrix.h" +#include "mex.h" +#include "TV_SB_GPU_core.h" + +/* CUDA mex-file for implementation of Split Bregman - TV denoising-regularisation model (2D/3D) [1] +* +* Input Parameters: +* 1. Noisy image/volume +* 2. lambda - regularisation parameter +* 3. Number of iterations [OPTIONAL parameter] +* 4. eplsilon - tolerance constant [OPTIONAL parameter] +* 5. TV-type: 'iso' or 'l1' [OPTIONAL parameter] +* 6. print information: 0 (off) or 1 (on) [OPTIONAL parameter] +* +* Output: +* 1. Filtered/regularized image +* +* This function is based on the Matlab's code and paper by +* [1]. Goldstein, T. and Osher, S., 2009. The split Bregman method for L1-regularized problems. SIAM journal on imaging sciences, 2(2), pp.323-343. +*/ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter, methTV, printswitch; + mwSize dimX, dimY, dimZ; + const mwSize *dim_array; + + float *Input, *Output=NULL, lambda, epsil; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + + /*Handling Matlab input data*/ + if ((nrhs < 2) || (nrhs > 6)) mexErrMsgTxt("At least 2 parameters is required, all parameters are: Image(2D/3D), Regularization parameter, Regularization parameter, iterations number, tolerance, penalty type ('iso' or 'l1'), print switch"); + + Input = (float *) mxGetData(prhs[0]); /*noisy image (2D/3D) */ + lambda = (float) mxGetScalar(prhs[1]); /* regularization parameter */ + iter = 100; /* default iterations number */ + epsil = 0.0001; /* default tolerance constant */ + methTV = 0; /* default isotropic TV penalty */ + printswitch = 0; /*default print is switched, off - 0 */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + + if ((nrhs == 3) || (nrhs == 4) || (nrhs == 5) || (nrhs == 6)) iter = (int) mxGetScalar(prhs[2]); /* iterations number */ + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6)) epsil = (float) mxGetScalar(prhs[3]); /* tolerance constant */ + if ((nrhs == 5) || (nrhs == 6)) { + char *penalty_type; + penalty_type = mxArrayToString(prhs[4]); /* choosing TV penalty: 'iso' or 'l1', 'iso' is the default */ + if ((strcmp(penalty_type, "l1") != 0) && (strcmp(penalty_type, "iso") != 0)) mexErrMsgTxt("Choose TV type: 'iso' or 'l1',"); + if (strcmp(penalty_type, "l1") == 0) methTV = 1; /* enable 'l1' penalty */ + mxFree(penalty_type); + } + if (nrhs == 6) { + printswitch = (int) mxGetScalar(prhs[5]); + if ((printswitch != 0) && (printswitch != 1)) mexErrMsgTxt("Print can be enabled by choosing 1 or off - 0"); + } + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2]; + + if (number_of_dims == 2) { + dimZ = 1; /*2D case*/ + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + } + if (number_of_dims == 3) Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL)); + + /* running the function */ + TV_SB_GPU_main(Input, Output, lambda, iter, epsil, methTV, printswitch, dimX, dimY, dimZ); +} diff --git a/src/Matlab/mex_compile/regularisers_GPU/TGV_GPU.cpp b/src/Matlab/mex_compile/regularisers_GPU/TGV_GPU.cpp new file mode 100644 index 0000000..edb551d --- /dev/null +++ b/src/Matlab/mex_compile/regularisers_GPU/TGV_GPU.cpp @@ -0,0 +1,79 @@ +/* +This work is part of the Core Imaging Library developed by +Visual Analytics and Imaging System Group of the Science Technology +Facilities Council, STFC + +Copyright 2017 Daniil Kazantsev +Copyright 2017 Srikanth Nagella, Edoardo Pasca + +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at +http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. +*/ + +#include "mex.h" +#include "TGV_GPU_core.h" + +/* CUDA implementation of Primal-Dual denoising method for + * Total Generilized Variation (TGV)-L2 model [1] (2D case only) + * + * Input Parameters: + * 1. Noisy image (2D) (required) + * 2. lambda - regularisation parameter (required) + * 3. parameter to control the first-order term (alpha1) (default - 1) + * 4. parameter to control the second-order term (alpha0) (default - 0.5) + * 5. Number of Chambolle-Pock (Primal-Dual) iterations (default is 300) + * 6. Lipshitz constant (default is 12) + * + * Output: + * Filtered/regulariaed image + * + * References: + * [1] K. Bredies "Total Generalized Variation" + */ + +void mexFunction( + int nlhs, mxArray *plhs[], + int nrhs, const mxArray *prhs[]) + +{ + int number_of_dims, iter; + mwSize dimX, dimY; + const mwSize *dim_array; + float *Input, *Output=NULL, lambda, alpha0, alpha1, L2; + + number_of_dims = mxGetNumberOfDimensions(prhs[0]); + dim_array = mxGetDimensions(prhs[0]); + + /*Handling Matlab input data*/ + if ((nrhs < 2) || (nrhs > 6)) mexErrMsgTxt("At least 2 parameters is required, all parameters are: Image(2D), Regularisation parameter, alpha0, alpha1, iterations number, Lipshitz Constant"); + + Input = (float *) mxGetData(prhs[0]); /*noisy image (2D) */ + lambda = (float) mxGetScalar(prhs[1]); /* regularisation parameter */ + alpha1 = 1.0f; /* parameter to control the first-order term */ + alpha0 = 0.5f; /* parameter to control the second-order term */ + iter = 300; /* Iterations number */ + L2 = 12.0f; /* Lipshitz constant */ + + if (mxGetClassID(prhs[0]) != mxSINGLE_CLASS) {mexErrMsgTxt("The input image must be in a single precision"); } + if ((nrhs == 3) || (nrhs == 4) || (nrhs == 5) || (nrhs == 6)) alpha1 = (float) mxGetScalar(prhs[2]); /* parameter to control the first-order term */ + if ((nrhs == 4) || (nrhs == 5) || (nrhs == 6)) alpha0 = (float) mxGetScalar(prhs[3]); /* parameter to control the second-order term */ + if ((nrhs == 5) || (nrhs == 6)) iter = (int) mxGetScalar(prhs[4]); /* Iterations number */ + if (nrhs == 6) L2 = (float) mxGetScalar(prhs[5]); /* Lipshitz constant */ + + /*Handling Matlab output data*/ + dimX = dim_array[0]; dimY = dim_array[1]; + + if (number_of_dims == 2) { + Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(2, dim_array, mxSINGLE_CLASS, mxREAL)); + /* running the function */ + TGV_GPU_main(Input, Output, lambda, alpha1, alpha0, iter, L2, dimX, dimY); + } + if (number_of_dims == 3) {mexErrMsgTxt("Only 2D images accepted");} +} diff --git a/src/Matlab/supp/RMSE.m b/src/Matlab/supp/RMSE.m new file mode 100644 index 0000000..002f776 --- /dev/null +++ b/src/Matlab/supp/RMSE.m @@ -0,0 +1,7 @@ +function err = RMSE(signal1, signal2)
+%RMSE Root Mean Squared Error
+
+err = sum((signal1 - signal2).^2)/length(signal1); % MSE
+err = sqrt(err); % RMSE
+
+end
\ No newline at end of file diff --git a/src/Matlab/supp/my_red_yellowMAP.mat b/src/Matlab/supp/my_red_yellowMAP.mat Binary files differnew file mode 100644 index 0000000..c2a5b87 --- /dev/null +++ b/src/Matlab/supp/my_red_yellowMAP.mat diff --git a/src/Python/CMakeLists.txt b/src/Python/CMakeLists.txt new file mode 100644 index 0000000..c2ef855 --- /dev/null +++ b/src/Python/CMakeLists.txt @@ -0,0 +1,141 @@ +# Copyright 2018 Edoardo Pasca +cmake_minimum_required (VERSION 3.0) + +project(regulariserPython) +#https://stackoverflow.com/questions/13298504/using-cmake-with-setup-py + +# The version number. + +#set (CIL_VERSION $ENV{CIL_VERSION} CACHE INTERNAL "Core Imaging Library version" FORCE) + +# conda orchestrated build +message("CIL_VERSION: ${CIL_VERSION}") +#include (GenerateExportHeader) + +find_package(PythonInterp REQUIRED) +if (PYTHONINTERP_FOUND) + message ("Current Python " ${PYTHON_VERSION_STRING} " found " ${PYTHON_EXECUTABLE}) +endif() + + +## Build the regularisers package as a library +message("Creating Regularisers as shared library") + +message("CMAKE_CXX_FLAGS ${CMAKE_CXX_FLAGS}") + +set(CMAKE_BUILD_TYPE "Release") + +if(WIN32) + set (FLAGS "/DWIN32 /EHsc /openmp /DCCPiCore_EXPORTS") + set (CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} /NODEFAULTLIB:MSVCRT.lib") + + set (EXTRA_LIBRARIES) + + message("library lib: ${LIBRARY_LIB}") + +elseif(UNIX) + set (FLAGS "-fopenmp -O2 -funsigned-char -Wall -Wl,--no-undefined -DCCPiReconstructionIterative_EXPORTS -std=c++0x") + set (EXTRA_LIBRARIES + "gomp" + ) +endif() + +# GPU regularisers +if (BUILD_CUDA) + find_package(CUDA) + if (CUDA_FOUND) + message("CUDA FOUND") + set (SETUP_GPU_WRAPPERS "extra_libraries += ['cilregcuda']\n\ +setup( \n\ + name='ccpi', \n\ + description='CCPi Core Imaging Library - Image regularisers GPU',\n\ + version=cil_version,\n\ + cmdclass = {'build_ext': build_ext},\n\ + ext_modules = [Extension('ccpi.filters.gpu_regularisers',\n\ + sources=[ \n\ + os.path.join('.' , 'src', 'gpu_regularisers.pyx' ),\n\ + ],\n\ + include_dirs=extra_include_dirs, \n\ + library_dirs=extra_library_dirs, \n\ + extra_compile_args=extra_compile_args, \n\ + libraries=extra_libraries ), \n\ + ],\n\ + zip_safe = False, \n\ + packages = {'ccpi','ccpi.filters'},\n\ + )") + else() + message("CUDA NOT FOUND") + set(SETUP_GPU_WRAPPERS "#CUDA NOT FOUND") + endif() +endif() +configure_file("${CMAKE_CURRENT_SOURCE_DIR}/setup-regularisers.py.in" "${CMAKE_CURRENT_BINARY_DIR}/setup-regularisers.py") + + +find_package(PythonInterp) +find_package(PythonLibs) +if (PYTHONINTERP_FOUND) + message(STATUS "Found PYTHON_EXECUTABLE=${PYTHON_EXECUTABLE}") + message(STATUS "Python version ${PYTHON_VERSION_STRING}") +endif() +if (PYTHONLIBS_FOUND) + message(STATUS "Found PYTHON_INCLUDE_DIRS=${PYTHON_INCLUDE_DIRS}") + message(STATUS "Found PYTHON_LIBRARIES=${PYTHON_LIBRARIES}") +endif() + +if (PYTHONINTERP_FOUND) + message("Python found " ${PYTHON_EXECUTABLE}) + set(SETUP_PY_IN "${CMAKE_CURRENT_SOURCE_DIR}/setup-regularisers.py.in") + set(SETUP_PY "${CMAKE_CURRENT_BINARY_DIR}/setup-regularisers.py") + #set(DEPS "${CMAKE_CURRENT_SOURCE_DIR}/module/__init__.py") + set (DEPS "${CMAKE_BINARY_DIR}/Core/") + set(OUTPUT "${CMAKE_CURRENT_BINARY_DIR}/build/timestamp") + + configure_file(${SETUP_PY_IN} ${SETUP_PY}) + + message("Core binary dir " ${CMAKE_BINARY_DIR}/Core/${CMAKE_BUILD_TYPE}) + + if (CONDA_BUILD) + add_custom_command(OUTPUT ${OUTPUT} + COMMAND ${CMAKE_COMMAND} -E copy_directory ${CMAKE_CURRENT_SOURCE_DIR}/src ${CMAKE_CURRENT_BINARY_DIR}/src + COMMAND ${CMAKE_COMMAND} -E copy_directory ${CMAKE_CURRENT_SOURCE_DIR}/ccpi ${CMAKE_CURRENT_BINARY_DIR}/ccpi + COMMAND ${CMAKE_COMMAND} -E env CIL_VERSION=${CIL_VERSION} + PREFIX=${CMAKE_SOURCE_DIR}/Core + LIBRARY_INC=${CMAKE_SOURCE_DIR}/Core + LIBRARY_LIB=${CMAKE_BINARY_DIR}/Core + ${PYTHON_EXECUTABLE} ${SETUP_PY} install + COMMAND ${CMAKE_COMMAND} -E touch ${OUTPUT} + DEPENDS cilreg) + + else() + if (WIN32) + add_custom_command(OUTPUT ${OUTPUT} + COMMAND ${CMAKE_COMMAND} -E copy_directory ${CMAKE_CURRENT_SOURCE_DIR}/src ${CMAKE_CURRENT_BINARY_DIR}/src + COMMAND ${CMAKE_COMMAND} -E copy_directory ${CMAKE_CURRENT_SOURCE_DIR}/ccpi ${CMAKE_CURRENT_BINARY_DIR}/ccpi + COMMAND ${CMAKE_COMMAND} -E env CIL_VERSION=${CIL_VERSION} + PREFIX=${CMAKE_SOURCE_DIR}/Core + LIBRARY_INC=${CMAKE_SOURCE_DIR}/Core + LIBRARY_LIB=${CMAKE_BINARY_DIR}/Core/${CMAKE_BUILD_TYPE} + ${PYTHON_EXECUTABLE} ${SETUP_PY} build_ext --inplace + COMMAND ${CMAKE_COMMAND} -E touch ${OUTPUT} + DEPENDS cilreg) + else() + add_custom_command(OUTPUT ${OUTPUT} + COMMAND ${CMAKE_COMMAND} -E copy_directory ${CMAKE_CURRENT_SOURCE_DIR}/src ${CMAKE_CURRENT_BINARY_DIR}/src + COMMAND ${CMAKE_COMMAND} -E copy_directory ${CMAKE_CURRENT_SOURCE_DIR}/ccpi ${CMAKE_CURRENT_BINARY_DIR}/ccpi + COMMAND ${CMAKE_COMMAND} -E env CIL_VERSION=${CIL_VERSION} + PREFIX=${CMAKE_SOURCE_DIR}/Core + LIBRARY_INC=${CMAKE_SOURCE_DIR}/Core + LIBRARY_LIB=${CMAKE_BINARY_DIR}/Core + ${PYTHON_EXECUTABLE} ${SETUP_PY} build_ext --inplace + COMMAND ${CMAKE_COMMAND} -E touch ${OUTPUT} + DEPENDS cilreg) + endif() + install(DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/ccpi + DESTINATION ${PYTHON_DEST}) + endif() + + + add_custom_target(PythonWrapper ALL DEPENDS ${OUTPUT}) + + #install(CODE "execute_process(COMMAND ${PYTHON} ${SETUP_PY} install)") +endif() diff --git a/src/Python/ccpi/__init__.py b/src/Python/ccpi/__init__.py new file mode 100644 index 0000000..e69de29 --- /dev/null +++ b/src/Python/ccpi/__init__.py diff --git a/src/Python/ccpi/filters/__init__.py b/src/Python/ccpi/filters/__init__.py new file mode 100644 index 0000000..e69de29 --- /dev/null +++ b/src/Python/ccpi/filters/__init__.py diff --git a/src/Python/ccpi/filters/regularisers.py b/src/Python/ccpi/filters/regularisers.py new file mode 100644 index 0000000..588ea32 --- /dev/null +++ b/src/Python/ccpi/filters/regularisers.py @@ -0,0 +1,214 @@ +""" +script which assigns a proper device core function based on a flag ('cpu' or 'gpu') +""" + +from ccpi.filters.cpu_regularisers import TV_ROF_CPU, TV_FGP_CPU, TV_SB_CPU, dTV_FGP_CPU, TNV_CPU, NDF_CPU, Diff4th_CPU, TGV_CPU, LLT_ROF_CPU, PATCHSEL_CPU, NLTV_CPU +try: + from ccpi.filters.gpu_regularisers import TV_ROF_GPU, TV_FGP_GPU, TV_SB_GPU, dTV_FGP_GPU, NDF_GPU, Diff4th_GPU, TGV_GPU, LLT_ROF_GPU, PATCHSEL_GPU + gpu_enabled = True +except ImportError: + gpu_enabled = False +from ccpi.filters.cpu_regularisers import NDF_INPAINT_CPU, NVM_INPAINT_CPU + +def ROF_TV(inputData, regularisation_parameter, iterations, + time_marching_parameter,device='cpu'): + if device == 'cpu': + return TV_ROF_CPU(inputData, + regularisation_parameter, + iterations, + time_marching_parameter) + elif device == 'gpu' and gpu_enabled: + return TV_ROF_GPU(inputData, + regularisation_parameter, + iterations, + time_marching_parameter) + else: + if not gpu_enabled and device == 'gpu': + raise ValueError ('GPU is not available') + raise ValueError('Unknown device {0}. Expecting gpu or cpu'\ + .format(device)) + +def FGP_TV(inputData, regularisation_parameter,iterations, + tolerance_param, methodTV, nonneg, printM, device='cpu'): + if device == 'cpu': + return TV_FGP_CPU(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + nonneg, + printM) + elif device == 'gpu' and gpu_enabled: + return TV_FGP_GPU(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + nonneg, + printM) + else: + if not gpu_enabled and device == 'gpu': + raise ValueError ('GPU is not available') + raise ValueError('Unknown device {0}. Expecting gpu or cpu'\ + .format(device)) +def SB_TV(inputData, regularisation_parameter, iterations, + tolerance_param, methodTV, printM, device='cpu'): + if device == 'cpu': + return TV_SB_CPU(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + printM) + elif device == 'gpu' and gpu_enabled: + return TV_SB_GPU(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + printM) + else: + if not gpu_enabled and device == 'gpu': + raise ValueError ('GPU is not available') + raise ValueError('Unknown device {0}. Expecting gpu or cpu'\ + .format(device)) +def FGP_dTV(inputData, refdata, regularisation_parameter, iterations, + tolerance_param, eta_const, methodTV, nonneg, printM, device='cpu'): + if device == 'cpu': + return dTV_FGP_CPU(inputData, + refdata, + regularisation_parameter, + iterations, + tolerance_param, + eta_const, + methodTV, + nonneg, + printM) + elif device == 'gpu' and gpu_enabled: + return dTV_FGP_GPU(inputData, + refdata, + regularisation_parameter, + iterations, + tolerance_param, + eta_const, + methodTV, + nonneg, + printM) + else: + if not gpu_enabled and device == 'gpu': + raise ValueError ('GPU is not available') + raise ValueError('Unknown device {0}. Expecting gpu or cpu'\ + .format(device)) +def TNV(inputData, regularisation_parameter, iterations, tolerance_param): + return TNV_CPU(inputData, + regularisation_parameter, + iterations, + tolerance_param) +def NDF(inputData, regularisation_parameter, edge_parameter, iterations, + time_marching_parameter, penalty_type, device='cpu'): + if device == 'cpu': + return NDF_CPU(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter, + penalty_type) + elif device == 'gpu' and gpu_enabled: + return NDF_GPU(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter, + penalty_type) + else: + if not gpu_enabled and device == 'gpu': + raise ValueError ('GPU is not available') + raise ValueError('Unknown device {0}. Expecting gpu or cpu'\ + .format(device)) +def Diff4th(inputData, regularisation_parameter, edge_parameter, iterations, + time_marching_parameter, device='cpu'): + if device == 'cpu': + return Diff4th_CPU(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter) + elif device == 'gpu' and gpu_enabled: + return Diff4th_GPU(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter) + else: + if not gpu_enabled and device == 'gpu': + raise ValueError ('GPU is not available') + raise ValueError('Unknown device {0}. Expecting gpu or cpu'\ + .format(device)) + +def PatchSelect(inputData, searchwindow, patchwindow, neighbours, edge_parameter, device='cpu'): + if device == 'cpu': + return PATCHSEL_CPU(inputData, + searchwindow, + patchwindow, + neighbours, + edge_parameter) + elif device == 'gpu' and gpu_enabled: + return PATCHSEL_GPU(inputData, + searchwindow, + patchwindow, + neighbours, + edge_parameter) + else: + if not gpu_enabled and device == 'gpu': + raise ValueError ('GPU is not available') + raise ValueError('Unknown device {0}. Expecting gpu or cpu'\ + .format(device)) + +def NLTV(inputData, H_i, H_j, H_k, Weights, regularisation_parameter, iterations): + return NLTV_CPU(inputData, + H_i, + H_j, + H_k, + Weights, + regularisation_parameter, + iterations) + +def TGV(inputData, regularisation_parameter, alpha1, alpha0, iterations, + LipshitzConst, device='cpu'): + if device == 'cpu': + return TGV_CPU(inputData, + regularisation_parameter, + alpha1, + alpha0, + iterations, + LipshitzConst) + elif device == 'gpu' and gpu_enabled: + return TGV_GPU(inputData, + regularisation_parameter, + alpha1, + alpha0, + iterations, + LipshitzConst) + else: + if not gpu_enabled and device == 'gpu': + raise ValueError ('GPU is not available') + raise ValueError('Unknown device {0}. Expecting gpu or cpu'\ + .format(device)) +def LLT_ROF(inputData, regularisation_parameterROF, regularisation_parameterLLT, iterations, + time_marching_parameter, device='cpu'): + if device == 'cpu': + return LLT_ROF_CPU(inputData, regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter) + elif device == 'gpu' and gpu_enabled: + return LLT_ROF_GPU(inputData, regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter) + else: + if not gpu_enabled and device == 'gpu': + raise ValueError ('GPU is not available') + raise ValueError('Unknown device {0}. Expecting gpu or cpu'\ + .format(device)) +def NDF_INP(inputData, maskData, regularisation_parameter, edge_parameter, iterations, + time_marching_parameter, penalty_type): + return NDF_INPAINT_CPU(inputData, maskData, regularisation_parameter, + edge_parameter, iterations, time_marching_parameter, penalty_type) + +def NVM_INP(inputData, maskData, SW_increment, iterations): + return NVM_INPAINT_CPU(inputData, maskData, SW_increment, iterations) diff --git a/src/Python/setup-regularisers.py.in b/src/Python/setup-regularisers.py.in new file mode 100644 index 0000000..462edda --- /dev/null +++ b/src/Python/setup-regularisers.py.in @@ -0,0 +1,75 @@ +#!/usr/bin/env python + +import setuptools +from distutils.core import setup +from distutils.extension import Extension +from Cython.Distutils import build_ext + +import os +import sys +import numpy +import platform + +cil_version=os.environ['CIL_VERSION'] +if cil_version == '': + print("Please set the environmental variable CIL_VERSION") + sys.exit(1) + +library_include_path = "" +library_lib_path = "" +try: + library_include_path = os.environ['LIBRARY_INC'] + library_lib_path = os.environ['LIBRARY_LIB'] +except: + library_include_path = os.environ['PREFIX']+'/include' + pass + +extra_include_dirs = [numpy.get_include(), library_include_path] +#extra_library_dirs = [os.path.join(library_include_path, "..", "lib")] +extra_compile_args = [] +extra_library_dirs = [library_lib_path] +extra_compile_args = [] +extra_link_args = [] +extra_libraries = ['cilreg'] + +print ("extra_library_dirs " , extra_library_dirs) + +extra_include_dirs += [os.path.join(".." , ".." , "Core"), + os.path.join(".." , ".." , "Core", "regularisers_CPU"), + os.path.join(".." , ".." , "Core", "inpainters_CPU"), + os.path.join(".." , ".." , "Core", "regularisers_GPU" , "TV_FGP" ) , + os.path.join(".." , ".." , "Core", "regularisers_GPU" , "TV_ROF" ) , + os.path.join(".." , ".." , "Core", "regularisers_GPU" , "TV_SB" ) , + os.path.join(".." , ".." , "Core", "regularisers_GPU" , "TGV" ) , + os.path.join(".." , ".." , "Core", "regularisers_GPU" , "LLTROF" ) , + os.path.join(".." , ".." , "Core", "regularisers_GPU" , "NDF" ) , + os.path.join(".." , ".." , "Core", "regularisers_GPU" , "dTV_FGP" ) , + os.path.join(".." , ".." , "Core", "regularisers_GPU" , "DIFF4th" ) , + os.path.join(".." , ".." , "Core", "regularisers_GPU" , "PatchSelect" ) , + "."] + +if platform.system() == 'Windows': + extra_compile_args[0:] = ['/DWIN32','/EHsc','/DBOOST_ALL_NO_LIB' , '/openmp' ] +else: + extra_compile_args = ['-fopenmp','-O2', '-funsigned-char', '-Wall', '-std=c++0x'] + extra_libraries += [@EXTRA_OMP_LIB@] + +setup( + name='ccpi', + description='CCPi Core Imaging Library - Image regularisers', + version=cil_version, + cmdclass = {'build_ext': build_ext}, + ext_modules = [Extension("ccpi.filters.cpu_regularisers", + sources=[os.path.join("." , "src", "cpu_regularisers.pyx" ) ], + include_dirs=extra_include_dirs, + library_dirs=extra_library_dirs, + extra_compile_args=extra_compile_args, + libraries=extra_libraries ), + + ], + zip_safe = False, + packages = {'ccpi','ccpi.filters'}, +) + + +@SETUP_GPU_WRAPPERS@ diff --git a/src/Python/src/cpu_regularisers.pyx b/src/Python/src/cpu_regularisers.pyx new file mode 100644 index 0000000..11a0617 --- /dev/null +++ b/src/Python/src/cpu_regularisers.pyx @@ -0,0 +1,685 @@ +# distutils: language=c++ +""" +Copyright 2018 CCPi +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at + http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. + +Author: Edoardo Pasca, Daniil Kazantsev +""" + +import cython +import numpy as np +cimport numpy as np + +cdef extern float TV_ROF_CPU_main(float *Input, float *Output, float lambdaPar, int iterationsNumb, float tau, int dimX, int dimY, int dimZ); +cdef extern float TV_FGP_CPU_main(float *Input, float *Output, float lambdaPar, int iterationsNumb, float epsil, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ); +cdef extern float SB_TV_CPU_main(float *Input, float *Output, float lambdaPar, int iterationsNumb, float epsil, int methodTV, int printM, int dimX, int dimY, int dimZ); +cdef extern float LLT_ROF_CPU_main(float *Input, float *Output, float lambdaROF, float lambdaLLT, int iterationsNumb, float tau, int dimX, int dimY, int dimZ); +cdef extern float TGV_main(float *Input, float *Output, float lambdaPar, float alpha1, float alpha0, int iterationsNumb, float L2, int dimX, int dimY, int dimZ); +cdef extern float Diffusion_CPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int penaltytype, int dimX, int dimY, int dimZ); +cdef extern float Diffus4th_CPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int dimX, int dimY, int dimZ); +cdef extern float TNV_CPU_main(float *Input, float *u, float lambdaPar, int maxIter, float tol, int dimX, int dimY, int dimZ); +cdef extern float dTV_FGP_CPU_main(float *Input, float *InputRef, float *Output, float lambdaPar, int iterationsNumb, float epsil, float eta, int methodTV, int nonneg, int printM, int dimX, int dimY, int dimZ); +cdef extern float PatchSelect_CPU_main(float *Input, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, int dimX, int dimY, int dimZ, int SearchWindow, int SimilarWin, int NumNeighb, float h, int switchM); +cdef extern float Nonlocal_TV_CPU_main(float *A_orig, float *Output, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, int dimX, int dimY, int dimZ, int NumNeighb, float lambdaReg, int IterNumb); + +cdef extern float Diffusion_Inpaint_CPU_main(float *Input, unsigned char *Mask, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int penaltytype, int dimX, int dimY, int dimZ); +cdef extern float NonlocalMarching_Inpaint_main(float *Input, unsigned char *M, float *Output, unsigned char *M_upd, int SW_increment, int iterationsNumb, int trigger, int dimX, int dimY, int dimZ); +cdef extern float TV_energy2D(float *U, float *U0, float *E_val, float lambdaPar, int type, int dimX, int dimY); +cdef extern float TV_energy3D(float *U, float *U0, float *E_val, float lambdaPar, int type, int dimX, int dimY, int dimZ); +#****************************************************************# +#********************** Total-variation ROF *********************# +#****************************************************************# +def TV_ROF_CPU(inputData, regularisation_parameter, iterationsNumb, marching_step_parameter): + if inputData.ndim == 2: + return TV_ROF_2D(inputData, regularisation_parameter, iterationsNumb, marching_step_parameter) + elif inputData.ndim == 3: + return TV_ROF_3D(inputData, regularisation_parameter, iterationsNumb, marching_step_parameter) + +def TV_ROF_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + int iterationsNumb, + float marching_step_parameter): + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Run ROF iterations for 2D data + TV_ROF_CPU_main(&inputData[0,0], &outputData[0,0], regularisation_parameter, iterationsNumb, marching_step_parameter, dims[1], dims[0], 1) + + return outputData + +def TV_ROF_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + int iterationsNumb, + float marching_step_parameter): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Run ROF iterations for 3D data + TV_ROF_CPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, iterationsNumb, marching_step_parameter, dims[2], dims[1], dims[0]) + + return outputData + +#****************************************************************# +#********************** Total-variation FGP *********************# +#****************************************************************# +#******** Total-variation Fast-Gradient-Projection (FGP)*********# +def TV_FGP_CPU(inputData, regularisation_parameter, iterationsNumb, tolerance_param, methodTV, nonneg, printM): + if inputData.ndim == 2: + return TV_FGP_2D(inputData, regularisation_parameter, iterationsNumb, tolerance_param, methodTV, nonneg, printM) + elif inputData.ndim == 3: + return TV_FGP_3D(inputData, regularisation_parameter, iterationsNumb, tolerance_param, methodTV, nonneg, printM) + +def TV_FGP_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + int iterationsNumb, + float tolerance_param, + int methodTV, + int nonneg, + int printM): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + #/* Run FGP-TV iterations for 2D data */ + TV_FGP_CPU_main(&inputData[0,0], &outputData[0,0], regularisation_parameter, + iterationsNumb, + tolerance_param, + methodTV, + nonneg, + printM, + dims[1],dims[0],1) + + return outputData + +def TV_FGP_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + int iterationsNumb, + float tolerance_param, + int methodTV, + int nonneg, + int printM): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0], dims[1], dims[2]], dtype='float32') + + #/* Run FGP-TV iterations for 3D data */ + TV_FGP_CPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, + iterationsNumb, + tolerance_param, + methodTV, + nonneg, + printM, + dims[2], dims[1], dims[0]) + return outputData + +#***************************************************************# +#********************** Total-variation SB *********************# +#***************************************************************# +#*************** Total-variation Split Bregman (SB)*************# +def TV_SB_CPU(inputData, regularisation_parameter, iterationsNumb, tolerance_param, methodTV, printM): + if inputData.ndim == 2: + return TV_SB_2D(inputData, regularisation_parameter, iterationsNumb, tolerance_param, methodTV, printM) + elif inputData.ndim == 3: + return TV_SB_3D(inputData, regularisation_parameter, iterationsNumb, tolerance_param, methodTV, printM) + +def TV_SB_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + int iterationsNumb, + float tolerance_param, + int methodTV, + int printM): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + #/* Run SB-TV iterations for 2D data */ + SB_TV_CPU_main(&inputData[0,0], &outputData[0,0], regularisation_parameter, + iterationsNumb, + tolerance_param, + methodTV, + printM, + dims[1],dims[0],1) + + return outputData + +def TV_SB_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + int iterationsNumb, + float tolerance_param, + int methodTV, + int printM): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0], dims[1], dims[2]], dtype='float32') + + #/* Run SB-TV iterations for 3D data */ + SB_TV_CPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, + iterationsNumb, + tolerance_param, + methodTV, + printM, + dims[2], dims[1], dims[0]) + return outputData + +#***************************************************************# +#***************** Total Generalised Variation *****************# +#***************************************************************# +def TGV_CPU(inputData, regularisation_parameter, alpha1, alpha0, iterations, LipshitzConst): + if inputData.ndim == 2: + return TGV_2D(inputData, regularisation_parameter, alpha1, alpha0, + iterations, LipshitzConst) + elif inputData.ndim == 3: + return TGV_3D(inputData, regularisation_parameter, alpha1, alpha0, + iterations, LipshitzConst) + +def TGV_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + float alpha1, + float alpha0, + int iterationsNumb, + float LipshitzConst): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + #/* Run TGV iterations for 2D data */ + TGV_main(&inputData[0,0], &outputData[0,0], regularisation_parameter, + alpha1, + alpha0, + iterationsNumb, + LipshitzConst, + dims[1],dims[0],1) + return outputData +def TGV_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + float alpha1, + float alpha0, + int iterationsNumb, + float LipshitzConst): + + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0], dims[1], dims[2]], dtype='float32') + + #/* Run TGV iterations for 3D data */ + TGV_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, + alpha1, + alpha0, + iterationsNumb, + LipshitzConst, + dims[2], dims[1], dims[0]) + return outputData + +#***************************************************************# +#******************* ROF - LLT regularisation ******************# +#***************************************************************# +def LLT_ROF_CPU(inputData, regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter): + if inputData.ndim == 2: + return LLT_ROF_2D(inputData, regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter) + elif inputData.ndim == 3: + return LLT_ROF_3D(inputData, regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter) + +def LLT_ROF_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameterROF, + float regularisation_parameterLLT, + int iterations, + float time_marching_parameter): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + #/* Run ROF-LLT iterations for 2D data */ + LLT_ROF_CPU_main(&inputData[0,0], &outputData[0,0], regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter, dims[1],dims[0],1) + return outputData + +def LLT_ROF_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameterROF, + float regularisation_parameterLLT, + int iterations, + float time_marching_parameter): + + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0], dims[1], dims[2]], dtype='float32') + + #/* Run ROF-LLT iterations for 3D data */ + LLT_ROF_CPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter, dims[2], dims[1], dims[0]) + return outputData + +#****************************************************************# +#**************Directional Total-variation FGP ******************# +#****************************************************************# +#******** Directional TV Fast-Gradient-Projection (FGP)*********# +def dTV_FGP_CPU(inputData, refdata, regularisation_parameter, iterationsNumb, tolerance_param, eta_const, methodTV, nonneg, printM): + if inputData.ndim == 2: + return dTV_FGP_2D(inputData, refdata, regularisation_parameter, iterationsNumb, tolerance_param, eta_const, methodTV, nonneg, printM) + elif inputData.ndim == 3: + return dTV_FGP_3D(inputData, refdata, regularisation_parameter, iterationsNumb, tolerance_param, eta_const, methodTV, nonneg, printM) + +def dTV_FGP_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + np.ndarray[np.float32_t, ndim=2, mode="c"] refdata, + float regularisation_parameter, + int iterationsNumb, + float tolerance_param, + float eta_const, + int methodTV, + int nonneg, + int printM): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + #/* Run FGP-dTV iterations for 2D data */ + dTV_FGP_CPU_main(&inputData[0,0], &refdata[0,0], &outputData[0,0], regularisation_parameter, + iterationsNumb, + tolerance_param, + eta_const, + methodTV, + nonneg, + printM, + dims[1], dims[0], 1) + + return outputData + +def dTV_FGP_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + np.ndarray[np.float32_t, ndim=3, mode="c"] refdata, + float regularisation_parameter, + int iterationsNumb, + float tolerance_param, + float eta_const, + int methodTV, + int nonneg, + int printM): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0], dims[1], dims[2]], dtype='float32') + + #/* Run FGP-dTV iterations for 3D data */ + dTV_FGP_CPU_main(&inputData[0,0,0], &refdata[0,0,0], &outputData[0,0,0], regularisation_parameter, + iterationsNumb, + tolerance_param, + eta_const, + methodTV, + nonneg, + printM, + dims[2], dims[1], dims[0]) + return outputData + +#****************************************************************# +#*********************Total Nuclear Variation********************# +#****************************************************************# +def TNV_CPU(inputData, regularisation_parameter, iterationsNumb, tolerance_param): + if inputData.ndim == 2: + return + elif inputData.ndim == 3: + return TNV_3D(inputData, regularisation_parameter, iterationsNumb, tolerance_param) + +def TNV_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + int iterationsNumb, + float tolerance_param): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Run TNV iterations for 3D (X,Y,Channels) data + TNV_CPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, iterationsNumb, tolerance_param, dims[2], dims[1], dims[0]) + return outputData +#****************************************************************# +#***************Nonlinear (Isotropic) Diffusion******************# +#****************************************************************# +def NDF_CPU(inputData, regularisation_parameter, edge_parameter, iterationsNumb,time_marching_parameter, penalty_type): + if inputData.ndim == 2: + return NDF_2D(inputData, regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type) + elif inputData.ndim == 3: + return NDF_3D(inputData, regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type) + +def NDF_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter, + int penalty_type): + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Run Nonlinear Diffusion iterations for 2D data + Diffusion_CPU_main(&inputData[0,0], &outputData[0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type, dims[1], dims[0], 1) + return outputData + +def NDF_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter, + int penalty_type): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Run Nonlinear Diffusion iterations for 3D data + Diffusion_CPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type, dims[2], dims[1], dims[0]) + + return outputData + +#****************************************************************# +#*************Anisotropic Fourth-Order diffusion*****************# +#****************************************************************# +def Diff4th_CPU(inputData, regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter): + if inputData.ndim == 2: + return Diff4th_2D(inputData, regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter) + elif inputData.ndim == 3: + return Diff4th_3D(inputData, regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter) + +def Diff4th_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter): + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Run Anisotropic Fourth-Order diffusion for 2D data + Diffus4th_CPU_main(&inputData[0,0], &outputData[0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, dims[1], dims[0], 1) + return outputData + +def Diff4th_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Run Anisotropic Fourth-Order diffusion for 3D data + Diffus4th_CPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, dims[2], dims[1], dims[0]) + + return outputData + +#****************************************************************# +#***************Patch-based weights calculation******************# +#****************************************************************# +def PATCHSEL_CPU(inputData, searchwindow, patchwindow, neighbours, edge_parameter): + if inputData.ndim == 2: + return PatchSel_2D(inputData, searchwindow, patchwindow, neighbours, edge_parameter) + elif inputData.ndim == 3: + return 1 +def PatchSel_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + int searchwindow, + int patchwindow, + int neighbours, + float edge_parameter): + cdef long dims[3] + dims[0] = neighbours + dims[1] = inputData.shape[0] + dims[2] = inputData.shape[1] + + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] Weights = \ + np.zeros([dims[0], dims[1],dims[2]], dtype='float32') + + cdef np.ndarray[np.uint16_t, ndim=3, mode="c"] H_i = \ + np.zeros([dims[0], dims[1],dims[2]], dtype='uint16') + + cdef np.ndarray[np.uint16_t, ndim=3, mode="c"] H_j = \ + np.zeros([dims[0], dims[1],dims[2]], dtype='uint16') + + # Run patch-based weight selection function + PatchSelect_CPU_main(&inputData[0,0], &H_j[0,0,0], &H_i[0,0,0], &H_i[0,0,0], &Weights[0,0,0], dims[2], dims[1], 0, searchwindow, patchwindow, neighbours, edge_parameter, 1) + return H_i, H_j, Weights +""" +def PatchSel_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + int searchwindow, + int patchwindow, + int neighbours, + float edge_parameter): + cdef long dims[4] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + dims[3] = neighbours + + cdef np.ndarray[np.float32_t, ndim=4, mode="c"] Weights = \ + np.zeros([dims[3],dims[0],dims[1],dims[2]], dtype='float32') + + cdef np.ndarray[np.uint16_t, ndim=4, mode="c"] H_i = \ + np.zeros([dims[3],dims[0],dims[1],dims[2]], dtype='uint16') + + cdef np.ndarray[np.uint16_t, ndim=4, mode="c"] H_j = \ + np.zeros([dims[3],dims[0],dims[1],dims[2]], dtype='uint16') + + cdef np.ndarray[np.uint16_t, ndim=4, mode="c"] H_k = \ + np.zeros([dims[3],dims[0],dims[1],dims[2]], dtype='uint16') + + # Run patch-based weight selection function + PatchSelect_CPU_main(&inputData[0,0,0], &H_i[0,0,0,0], &H_j[0,0,0,0], &H_k[0,0,0,0], &Weights[0,0,0,0], dims[2], dims[1], dims[0], searchwindow, patchwindow, neighbours, edge_parameter, 1) + return H_i, H_j, H_k, Weights +""" + +#****************************************************************# +#***************Non-local Total Variation******************# +#****************************************************************# +def NLTV_CPU(inputData, H_i, H_j, H_k, Weights, regularisation_parameter, iterations): + if inputData.ndim == 2: + return NLTV_2D(inputData, H_i, H_j, Weights, regularisation_parameter, iterations) + elif inputData.ndim == 3: + return 1 +def NLTV_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + np.ndarray[np.uint16_t, ndim=3, mode="c"] H_i, + np.ndarray[np.uint16_t, ndim=3, mode="c"] H_j, + np.ndarray[np.float32_t, ndim=3, mode="c"] Weights, + float regularisation_parameter, + int iterations): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + neighbours = H_i.shape[0] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Run nonlocal TV regularisation + Nonlocal_TV_CPU_main(&inputData[0,0], &outputData[0,0], &H_i[0,0,0], &H_j[0,0,0], &H_i[0,0,0], &Weights[0,0,0], dims[1], dims[0], 0, neighbours, regularisation_parameter, iterations) + return outputData + +#*********************Inpainting WITH****************************# +#***************Nonlinear (Isotropic) Diffusion******************# +#****************************************************************# +def NDF_INPAINT_CPU(inputData, maskData, regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type): + if inputData.ndim == 2: + return NDF_INP_2D(inputData, maskData, regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type) + elif inputData.ndim == 3: + return NDF_INP_3D(inputData, maskData, regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type) + +def NDF_INP_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + np.ndarray[np.uint8_t, ndim=2, mode="c"] maskData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter, + int penalty_type): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Run Inpaiting by Diffusion iterations for 2D data + Diffusion_Inpaint_CPU_main(&inputData[0,0], &maskData[0,0], &outputData[0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type, dims[1], dims[0], 1) + return outputData + +def NDF_INP_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + np.ndarray[np.uint8_t, ndim=3, mode="c"] maskData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter, + int penalty_type): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Run Inpaiting by Diffusion iterations for 3D data + Diffusion_Inpaint_CPU_main(&inputData[0,0,0], &maskData[0,0,0], &outputData[0,0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type, dims[2], dims[1], dims[0]) + + return outputData +#*********************Inpainting WITH****************************# +#***************Nonlocal Vertical Marching method****************# +#****************************************************************# +def NVM_INPAINT_CPU(inputData, maskData, SW_increment, iterationsNumb): + if inputData.ndim == 2: + return NVM_INP_2D(inputData, maskData, SW_increment, iterationsNumb) + elif inputData.ndim == 3: + return + +def NVM_INP_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + np.ndarray[np.uint8_t, ndim=2, mode="c"] maskData, + int SW_increment, + int iterationsNumb): + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + cdef np.ndarray[np.uint8_t, ndim=2, mode="c"] maskData_upd = \ + np.zeros([dims[0],dims[1]], dtype='uint8') + + # Run Inpaiting by Nonlocal vertical marching method for 2D data + NonlocalMarching_Inpaint_main(&inputData[0,0], &maskData[0,0], &outputData[0,0], + &maskData_upd[0,0], + SW_increment, iterationsNumb, 1, dims[1], dims[0], 1) + + return (outputData, maskData_upd) + + +#****************************************************************# +#***************Calculation of TV-energy functional**************# +#****************************************************************# +def TV_ENERGY(inputData, inputData0, regularisation_parameter, typeFunctional): + if inputData.ndim == 2: + return TV_ENERGY_2D(inputData, inputData0, regularisation_parameter, typeFunctional) + elif inputData.ndim == 3: + return TV_ENERGY_3D(inputData, inputData0, regularisation_parameter, typeFunctional) + +def TV_ENERGY_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + np.ndarray[np.float32_t, ndim=2, mode="c"] inputData0, + float regularisation_parameter, + int typeFunctional): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=1, mode="c"] outputData = \ + np.zeros([1], dtype='float32') + + # run function + TV_energy2D(&inputData[0,0], &inputData0[0,0], &outputData[0], regularisation_parameter, typeFunctional, dims[1], dims[0]) + + return outputData + +def TV_ENERGY_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + np.ndarray[np.float32_t, ndim=3, mode="c"] inputData0, + float regularisation_parameter, + int typeFunctional): + + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=1, mode="c"] outputData = \ + np.zeros([1], dtype='float32') + + # Run function + TV_energy3D(&inputData[0,0,0], &inputData0[0,0,0], &outputData[0], regularisation_parameter, typeFunctional, dims[2], dims[1], dims[0]) + + return outputData diff --git a/src/Python/src/gpu_regularisers.pyx b/src/Python/src/gpu_regularisers.pyx new file mode 100644 index 0000000..b52f669 --- /dev/null +++ b/src/Python/src/gpu_regularisers.pyx @@ -0,0 +1,640 @@ +# distutils: language=c++ +""" +Copyright 2018 CCPi +Licensed under the Apache License, Version 2.0 (the "License"); +you may not use this file except in compliance with the License. +You may obtain a copy of the License at + http://www.apache.org/licenses/LICENSE-2.0 +Unless required by applicable law or agreed to in writing, software +distributed under the License is distributed on an "AS IS" BASIS, +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +See the License for the specific language governing permissions and +limitations under the License. + +Author: Edoardo Pasca, Daniil Kazantsev +""" + +import cython +import numpy as np +cimport numpy as np + +CUDAErrorMessage = 'CUDA error' + +cdef extern int TV_ROF_GPU_main(float* Input, float* Output, float lambdaPar, int iter, float tau, int N, int M, int Z); +cdef extern int TV_FGP_GPU_main(float *Input, float *Output, float lambdaPar, int iter, float epsil, int methodTV, int nonneg, int printM, int N, int M, int Z); +cdef extern int TV_SB_GPU_main(float *Input, float *Output, float lambdaPar, int iter, float epsil, int methodTV, int printM, int N, int M, int Z); +cdef extern int TGV_GPU_main(float *Input, float *Output, float lambdaPar, float alpha1, float alpha0, int iterationsNumb, float L2, int dimX, int dimY, int dimZ); +cdef extern int LLT_ROF_GPU_main(float *Input, float *Output, float lambdaROF, float lambdaLLT, int iterationsNumb, float tau, int N, int M, int Z); +cdef extern int NonlDiff_GPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int penaltytype, int N, int M, int Z); +cdef extern int dTV_FGP_GPU_main(float *Input, float *InputRef, float *Output, float lambdaPar, int iterationsNumb, float epsil, float eta, int methodTV, int nonneg, int printM, int N, int M, int Z); +cdef extern int Diffus4th_GPU_main(float *Input, float *Output, float lambdaPar, float sigmaPar, int iterationsNumb, float tau, int N, int M, int Z); +cdef extern int PatchSelect_GPU_main(float *Input, unsigned short *H_i, unsigned short *H_j, float *Weights, int N, int M, int SearchWindow, int SimilarWin, int NumNeighb, float h); + +# Total-variation Rudin-Osher-Fatemi (ROF) +def TV_ROF_GPU(inputData, + regularisation_parameter, + iterations, + time_marching_parameter): + if inputData.ndim == 2: + return ROFTV2D(inputData, + regularisation_parameter, + iterations, + time_marching_parameter) + elif inputData.ndim == 3: + return ROFTV3D(inputData, + regularisation_parameter, + iterations, + time_marching_parameter) + +# Total-variation Fast-Gradient-Projection (FGP) +def TV_FGP_GPU(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + nonneg, + printM): + if inputData.ndim == 2: + return FGPTV2D(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + nonneg, + printM) + elif inputData.ndim == 3: + return FGPTV3D(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + nonneg, + printM) +# Total-variation Split Bregman (SB) +def TV_SB_GPU(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + printM): + if inputData.ndim == 2: + return SBTV2D(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + printM) + elif inputData.ndim == 3: + return SBTV3D(inputData, + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + printM) +# LLT-ROF model +def LLT_ROF_GPU(inputData, regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter): + if inputData.ndim == 2: + return LLT_ROF_GPU2D(inputData, regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter) + elif inputData.ndim == 3: + return LLT_ROF_GPU3D(inputData, regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter) +# Total Generilised Variation (TGV) +def TGV_GPU(inputData, regularisation_parameter, alpha1, alpha0, iterations, LipshitzConst): + if inputData.ndim == 2: + return TGV2D(inputData, regularisation_parameter, alpha1, alpha0, iterations, LipshitzConst) + elif inputData.ndim == 3: + return TGV3D(inputData, regularisation_parameter, alpha1, alpha0, iterations, LipshitzConst) +# Directional Total-variation Fast-Gradient-Projection (FGP) +def dTV_FGP_GPU(inputData, + refdata, + regularisation_parameter, + iterations, + tolerance_param, + eta_const, + methodTV, + nonneg, + printM): + if inputData.ndim == 2: + return FGPdTV2D(inputData, + refdata, + regularisation_parameter, + iterations, + tolerance_param, + eta_const, + methodTV, + nonneg, + printM) + elif inputData.ndim == 3: + return FGPdTV3D(inputData, + refdata, + regularisation_parameter, + iterations, + tolerance_param, + eta_const, + methodTV, + nonneg, + printM) +# Nonlocal Isotropic Diffusion (NDF) +def NDF_GPU(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter, + penalty_type): + if inputData.ndim == 2: + return NDF_GPU_2D(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter, + penalty_type) + elif inputData.ndim == 3: + return NDF_GPU_3D(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter, + penalty_type) +# Anisotropic Fourth-Order diffusion +def Diff4th_GPU(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter): + if inputData.ndim == 2: + return Diff4th_2D(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter) + elif inputData.ndim == 3: + return Diff4th_3D(inputData, + regularisation_parameter, + edge_parameter, + iterations, + time_marching_parameter) + +#****************************************************************# +#********************** Total-variation ROF *********************# +#****************************************************************# +def ROFTV2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + int iterations, + float time_marching_parameter): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Running CUDA code here + if (TV_ROF_GPU_main( + &inputData[0,0], &outputData[0,0], + regularisation_parameter, + iterations , + time_marching_parameter, + dims[1], dims[0], 1)==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + +def ROFTV3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + int iterations, + float time_marching_parameter): + + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Running CUDA code here + if (TV_ROF_GPU_main( + &inputData[0,0,0], &outputData[0,0,0], + regularisation_parameter, + iterations , + time_marching_parameter, + dims[2], dims[1], dims[0])==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); +#****************************************************************# +#********************** Total-variation FGP *********************# +#****************************************************************# +#******** Total-variation Fast-Gradient-Projection (FGP)*********# +def FGPTV2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + int iterations, + float tolerance_param, + int methodTV, + int nonneg, + int printM): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Running CUDA code here + if (TV_FGP_GPU_main(&inputData[0,0], &outputData[0,0], + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + nonneg, + printM, + dims[1], dims[0], 1)==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + + +def FGPTV3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + int iterations, + float tolerance_param, + int methodTV, + int nonneg, + int printM): + + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Running CUDA code here + if (TV_FGP_GPU_main(&inputData[0,0,0], &outputData[0,0,0], + regularisation_parameter , + iterations, + tolerance_param, + methodTV, + nonneg, + printM, + dims[2], dims[1], dims[0])==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + +#***************************************************************# +#********************** Total-variation SB *********************# +#***************************************************************# +#*************** Total-variation Split Bregman (SB)*************# +def SBTV2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + int iterations, + float tolerance_param, + int methodTV, + int printM): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Running CUDA code here + if (TV_SB_GPU_main(&inputData[0,0], &outputData[0,0], + regularisation_parameter, + iterations, + tolerance_param, + methodTV, + printM, + dims[1], dims[0], 1)==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + + +def SBTV3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + int iterations, + float tolerance_param, + int methodTV, + int printM): + + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Running CUDA code here + if (TV_SB_GPU_main(&inputData[0,0,0], &outputData[0,0,0], + regularisation_parameter , + iterations, + tolerance_param, + methodTV, + printM, + dims[2], dims[1], dims[0])==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + + +#***************************************************************# +#************************ LLT-ROF model ************************# +#***************************************************************# +#************Joint LLT-ROF model for higher order **************# +def LLT_ROF_GPU2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameterROF, + float regularisation_parameterLLT, + int iterations, + float time_marching_parameter): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Running CUDA code here + if (LLT_ROF_GPU_main(&inputData[0,0], &outputData[0,0],regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter, dims[1],dims[0],1)==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + + +def LLT_ROF_GPU3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameterROF, + float regularisation_parameterLLT, + int iterations, + float time_marching_parameter): + + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Running CUDA code here + if (LLT_ROF_GPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameterROF, regularisation_parameterLLT, iterations, time_marching_parameter, dims[2], dims[1], dims[0])==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + + +#***************************************************************# +#***************** Total Generalised Variation *****************# +#***************************************************************# +def TGV2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + float alpha1, + float alpha0, + int iterationsNumb, + float LipshitzConst): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + #/* Run TGV iterations for 2D data */ + if (TGV_GPU_main(&inputData[0,0], &outputData[0,0], regularisation_parameter, + alpha1, + alpha0, + iterationsNumb, + LipshitzConst, + dims[1],dims[0], 1)==0): + return outputData + else: + raise ValueError(CUDAErrorMessage); + +def TGV3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + float alpha1, + float alpha0, + int iterationsNumb, + float LipshitzConst): + + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Running CUDA code here + if (TGV_GPU_main( + &inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, + alpha1, + alpha0, + iterationsNumb, + LipshitzConst, + dims[2], dims[1], dims[0])==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + + +#****************************************************************# +#**************Directional Total-variation FGP ******************# +#****************************************************************# +#******** Directional TV Fast-Gradient-Projection (FGP)*********# +def FGPdTV2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + np.ndarray[np.float32_t, ndim=2, mode="c"] refdata, + float regularisation_parameter, + int iterations, + float tolerance_param, + float eta_const, + int methodTV, + int nonneg, + int printM): + + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Running CUDA code here + if (dTV_FGP_GPU_main(&inputData[0,0], &refdata[0,0], &outputData[0,0], + regularisation_parameter, + iterations, + tolerance_param, + eta_const, + methodTV, + nonneg, + printM, + dims[1], dims[0], 1)==0): + return outputData + else: + raise ValueError(CUDAErrorMessage); + + +def FGPdTV3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + np.ndarray[np.float32_t, ndim=3, mode="c"] refdata, + float regularisation_parameter, + int iterations, + float tolerance_param, + float eta_const, + int methodTV, + int nonneg, + int printM): + + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Running CUDA code here + if (dTV_FGP_GPU_main(&inputData[0,0,0], &refdata[0,0,0], &outputData[0,0,0], + regularisation_parameter , + iterations, + tolerance_param, + eta_const, + methodTV, + nonneg, + printM, + dims[2], dims[1], dims[0])==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + + +#****************************************************************# +#***************Nonlinear (Isotropic) Diffusion******************# +#****************************************************************# +def NDF_GPU_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter, + int penalty_type): + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + #rangecheck = penalty_type < 1 and penalty_type > 3 + #if not rangecheck: +# raise ValueError('Choose penalty type as 1 for Huber, 2 - Perona-Malik, 3 - Tukey Biweight') + + # Run Nonlinear Diffusion iterations for 2D data + # Running CUDA code here + if (NonlDiff_GPU_main(&inputData[0,0], &outputData[0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type, dims[1], dims[0], 1)==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + + +def NDF_GPU_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter, + int penalty_type): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Run Nonlinear Diffusion iterations for 3D data + # Running CUDA code here + if (NonlDiff_GPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, penalty_type, dims[2], dims[1], dims[0])==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + +#****************************************************************# +#************Anisotropic Fourth-Order diffusion******************# +#****************************************************************# +def Diff4th_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter): + cdef long dims[2] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=2, mode="c"] outputData = \ + np.zeros([dims[0],dims[1]], dtype='float32') + + # Run Anisotropic Fourth-Order diffusion for 2D data + # Running CUDA code here + if (Diffus4th_GPU_main(&inputData[0,0], &outputData[0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, dims[1], dims[0], 1)==0): + return outputData + else: + raise ValueError(CUDAErrorMessage); + + +def Diff4th_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData, + float regularisation_parameter, + float edge_parameter, + int iterationsNumb, + float time_marching_parameter): + cdef long dims[3] + dims[0] = inputData.shape[0] + dims[1] = inputData.shape[1] + dims[2] = inputData.shape[2] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] outputData = \ + np.zeros([dims[0],dims[1],dims[2]], dtype='float32') + + # Run Anisotropic Fourth-Order diffusion for 3D data + # Running CUDA code here + if (Diffus4th_GPU_main(&inputData[0,0,0], &outputData[0,0,0], regularisation_parameter, edge_parameter, iterationsNumb, time_marching_parameter, dims[2], dims[1], dims[0])==0): + return outputData; + else: + raise ValueError(CUDAErrorMessage); + +#****************************************************************# +#************Patch-based weights pre-selection******************# +#****************************************************************# +def PATCHSEL_GPU(inputData, searchwindow, patchwindow, neighbours, edge_parameter): + if inputData.ndim == 2: + return PatchSel_2D(inputData, searchwindow, patchwindow, neighbours, edge_parameter) + elif inputData.ndim == 3: + return 1 +def PatchSel_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, + int searchwindow, + int patchwindow, + int neighbours, + float edge_parameter): + cdef long dims[3] + dims[0] = neighbours + dims[1] = inputData.shape[0] + dims[2] = inputData.shape[1] + + cdef np.ndarray[np.float32_t, ndim=3, mode="c"] Weights = \ + np.zeros([dims[0], dims[1],dims[2]], dtype='float32') + + cdef np.ndarray[np.uint16_t, ndim=3, mode="c"] H_i = \ + np.zeros([dims[0], dims[1],dims[2]], dtype='uint16') + + cdef np.ndarray[np.uint16_t, ndim=3, mode="c"] H_j = \ + np.zeros([dims[0], dims[1],dims[2]], dtype='uint16') + + # Run patch-based weight selection function + if (PatchSelect_GPU_main(&inputData[0,0], &H_j[0,0,0], &H_i[0,0,0], &Weights[0,0,0], dims[2], dims[1], searchwindow, patchwindow, neighbours, edge_parameter)==0): + return H_i, H_j, Weights; + else: + raise ValueError(CUDAErrorMessage); + diff --git a/test/lena_gray_512.tif b/test/lena_gray_512.tif Binary files differnew file mode 100644 index 0000000..f80cafc --- /dev/null +++ b/test/lena_gray_512.tif diff --git a/test/test_ROF_TV.py b/test/test_ROF_TV.py new file mode 100644 index 0000000..dda38b7 --- /dev/null +++ b/test/test_ROF_TV.py @@ -0,0 +1,127 @@ +import unittest +import math +import os +import timeit +from ccpi.filters.regularisers import ROF_TV +#, FGP_TV, SB_TV, TGV, LLT_ROF, FGP_dTV, NDF, Diff4th +from testroutines import * + +class TestRegularisers(unittest.TestCase): + + def test_ROF_TV_CPU(self): + filename = os.path.join("lena_gray_512.tif") + plt = TiffReader() + # read image + Im = plt.imread(filename) + Im = np.asarray(Im, dtype='float32') + + Im = Im / 255 + perc = 0.05 + u0 = Im + np.random.normal(loc=0, + scale=perc * Im, + size=np.shape(Im)) + u_ref = Im + np.random.normal(loc=0, + scale=0.01 * Im, + size=np.shape(Im)) + + # map the u0 u0->u0>0 + # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1) + u0 = u0.astype('float32') + u_ref = u_ref.astype('float32') + + + # set parameters + pars = {'algorithm': ROF_TV, \ + 'input': u0, \ + 'regularisation_parameter': 0.04, \ + 'number_of_iterations': 2500, \ + 'time_marching_parameter': 0.00002 + } + print("#############ROF TV CPU####################") + start_time = timeit.default_timer() + rof_cpu = ROF_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], 'cpu') + rms = rmse(Im, rof_cpu) + pars['rmse'] = rms + txtstr = printParametersToString(pars) + txtstr += "%s = %.3fs" % ('elapsed time', timeit.default_timer() - start_time) + print(txtstr) + + self.assertTrue(math.isclose(rms,0.02067839,rel_tol=1e-2)) + + + def test_ROF_TV_CPU_vs_GPU(self): + # print ("tomas debug test function") + print(__name__) + self.fail("testfail2") + filename = os.path.join("lena_gray_512.tif") + plt = TiffReader() + # read image + Im = plt.imread(filename) + Im = np.asarray(Im, dtype='float32') + + Im = Im / 255 + perc = 0.05 + u0 = Im + np.random.normal(loc=0, + scale=perc * Im, + size=np.shape(Im)) + u_ref = Im + np.random.normal(loc=0, + scale=0.01 * Im, + size=np.shape(Im)) + + # map the u0 u0->u0>0 + # f = np.frompyfunc(lambda x: 0 if x < 0 else x, 1,1) + u0 = u0.astype('float32') + u_ref = u_ref.astype('float32') + + print("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + print("____________ROF-TV bench___________________") + print("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%") + + # set parameters + pars = {'algorithm': ROF_TV, \ + 'input': u0, \ + 'regularisation_parameter': 0.04, \ + 'number_of_iterations': 2500, \ + 'time_marching_parameter': 0.00002 + } + print("##############ROF TV GPU##################") + start_time = timeit.default_timer() + try: + rof_gpu = ROF_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], 'gpu') + except ValueError as ve: + self.skipTest("Results not comparable. GPU computing error.") + + rms = rmse(Im, rof_gpu) + pars['rmse'] = rms + pars['algorithm'] = ROF_TV + txtstr = printParametersToString(pars) + txtstr += "%s = %.3fs" % ('elapsed time', timeit.default_timer() - start_time) + print(txtstr) + + print("#############ROF TV CPU####################") + start_time = timeit.default_timer() + rof_cpu = ROF_TV(pars['input'], + pars['regularisation_parameter'], + pars['number_of_iterations'], + pars['time_marching_parameter'], 'cpu') + rms = rmse(Im, rof_cpu) + pars['rmse'] = rms + + txtstr = printParametersToString(pars) + txtstr += "%s = %.3fs" % ('elapsed time', timeit.default_timer() - start_time) + print(txtstr) + print("--------Compare the results--------") + tolerance = 1e-04 + diff_im = np.zeros(np.shape(rof_cpu)) + diff_im = abs(rof_cpu - rof_gpu) + diff_im[diff_im > tolerance] = 1 + self.assertLessEqual(diff_im.sum(), 1) + +if __name__ == '__main__': + unittest.main() diff --git a/test/testroutines.py b/test/testroutines.py new file mode 100644 index 0000000..8da5c5e --- /dev/null +++ b/test/testroutines.py @@ -0,0 +1,37 @@ +import numpy as np +from PIL import Image + +class TiffReader(object): + def imread(self, filename): + return np.asarray(Image.open(filename)) + + +############################################################################### +def printParametersToString(pars): + txt = r'' + for key, value in pars.items(): + if key == 'algorithm': + txt += "{0} = {1}".format(key, value.__name__) + elif key == 'input': + txt += "{0} = {1}".format(key, np.shape(value)) + elif key == 'refdata': + txt += "{0} = {1}".format(key, np.shape(value)) + else: + txt += "{0} = {1}".format(key, value) + txt += '\n' + return txt + + +def nrmse(im1, im2): + rmse = np.sqrt(np.sum((im2 - im1) ** 2) / float(im1.size)) + max_val = max(np.max(im1), np.max(im2)) + min_val = min(np.min(im1), np.min(im2)) + return 1 - (rmse / (max_val - min_val)) + + +def rmse(im1, im2): + rmse = np.sqrt(np.sum((im1 - im2) ** 2) / float(im1.size)) + return rmse + + +############################################################################### |