cythonised nltv and updated demo, readme, bash run added

18 files changed, 327 insertions, 618 deletions

diff --git a/Core/regularisers_CPU/Nonlocal_TV_core.c b/Core/regularisers_CPU/Nonlocal_TV_core.c
index d327dd5..c4c9118 100644
--- a/Core/regularisers_CPU/Nonlocal_TV_core.c
+++ b/Core/regularisers_CPU/Nonlocal_TV_core.c
@@ -42,36 +42,37 @@
  */
 /*****************************************************************************/
 
-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 lambda, int IterNumb)
+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);       
+	  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, dimX, dimY, NumNeighb, lambda); */ /* NLM - H1 penalty */
-            NLM_TV_2D(Output, A_orig, H_i, H_j, Weights, i, j, (long)(dimX), (long)(dimY), NumNeighb, lambda);   /* NLM - TV penalty */     
-           }}          
+             /*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));       
+        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, lambda); */ /* 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, lambda);   /* NLM - TV penalty */     
+            /* 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 */     
            }}}          
           }          
     }
@@ -79,23 +80,24 @@ float Nonlocal_TV_CPU_main(float *A_orig, float *Output, unsigned short *H_i, un
 }
 
 /***********<<<<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 lambda)
+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; 
+	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[j*dimX+i] = (lambda*A_orig[j*dimX+i] + value)/(lambda + normweight);
+	}
+	 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 lambda)
+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;
@@ -108,39 +110,41 @@ float NLM_H1_3D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_
 		value += A[(dimX*dimY*k1) + j1*dimX+i1]*Weights[index];
 		normweight += Weights[index];
 	}	
-    A[(dimX*dimY*k) + j*dimX+i] = (lambda*A_orig[(dimX*dimY*k) + j*dimX+i] + value)/(lambda + normweight);
+    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 lambda)
+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; 
+	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;
+		index =  (dimX*dimY*x) + j*dimX+i; /*c*/
 		i1 = H_i[index];
 		j1 = H_j[index];
-	        NLgrad_magn += powf((A[j1*dimX+i1] - A[j*dimX+i]),2)*Weights[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;
+	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[j*dimX+i] = (lambda*A_orig[j*dimX+i] + value)/(lambda + normweight);
+    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 lambda)
+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;
@@ -164,6 +168,6 @@ float NLM_TV_3D(float *A, float *A_orig, unsigned short *H_i, unsigned short *H_
         value += A[(dimX*dimY*k1) + j1*dimX+i1]*NLCoeff*Weights[index];
         normweight += Weights[index]*NLCoeff;
     }   		
-    A[(dimX*dimY*k) + j*dimX+i] = (lambda*A_orig[(dimX*dimY*k) + j*dimX+i] + value)/(lambda + normweight);
+    A[(dimX*dimY*k) + j*dimX+i] = (lambdaReg*A_orig[(dimX*dimY*k) + j*dimX+i] + value)/(lambdaReg + normweight);
     return *A;
 }
diff --git a/Core/regularisers_CPU/Nonlocal_TV_core.h b/Core/regularisers_CPU/Nonlocal_TV_core.h
index 5b6963e..6d55101 100644
--- a/Core/regularisers_CPU/Nonlocal_TV_core.h
+++ b/Core/regularisers_CPU/Nonlocal_TV_core.h
@@ -51,11 +51,11 @@
 #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 lambda, 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 lambda);
-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 lambda);
-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 lambda);
-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 lambda);
+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/Core/regularisers_CPU/PatchSelect_core.c b/Core/regularisers_CPU/PatchSelect_core.c
index efc5498..45cf8c8 100644
--- a/Core/regularisers_CPU/PatchSelect_core.c
+++ b/Core/regularisers_CPU/PatchSelect_core.c
@@ -46,15 +46,15 @@
 
 /**************************************************/
 
-float PatchSelect_CPU_main(float *A, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, long dimX, long dimY, long dimZ, int SearchWindow, int SimilarWin, int NumNeighb, float h)
+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;   
+    h2 = h*h;
    
     /****************2D INPUT ***************/
-    if (dimZ == 0) {       
+    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;
@@ -63,13 +63,21 @@ float PatchSelect_CPU_main(float *A, unsigned short *H_i, unsigned short *H_j, u
                 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<dimX; i++) {
-            for(j=0; j<dimY; j++) {
-                Indeces2D(A, H_i, H_j, Weights, (i), (j), (dimX), (dimY), Eucl_Vec, NumNeighb, SearchWindow, SimilarWin, h2);
+    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 ***************/       
@@ -84,18 +92,28 @@ float PatchSelect_CPU_main(float *A, unsigned short *H_i, unsigned short *H_j, u
                 }}} /*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 dimY, long dimX, float *Eucl_Vec, int NumNeighb, int SearchWindow, int SimilarWin, float h2)
+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, temp;
@@ -167,6 +185,81 @@ float Indeces2D(float *Aorig, unsigned short *H_i, unsigned short *H_j, float *W
     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, temp;
+    unsigned short *ind_i, *ind_j, temp_i, temp_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; 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];
+                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;
+            }}}
+    /*sorting loop finished*/
+    
+    /*now select the NumNeighb more prominent weights and store into arrays */
+    for(x=0; x < NumNeighb; x++) {
+        //index = (dimX*dimY*x) + j*dimX+i;
+        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;
diff --git a/Core/regularisers_CPU/PatchSelect_core.h b/Core/regularisers_CPU/PatchSelect_core.h
index 43fce87..ddaa428 100644
--- a/Core/regularisers_CPU/PatchSelect_core.h
+++ b/Core/regularisers_CPU/PatchSelect_core.h
@@ -54,8 +54,9 @@
 #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, long dimX, long dimY, long dimZ, int SearchWindow, int SimilarWin, int NumNeighb, float h);
-CCPI_EXPORT float Indeces2D(float *Aorig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimY, long dimX, float *Eucl_Vec, int NumNeighb, int SearchWindow, int SimilarWin, float h2);
+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
 }
diff --git a/Readme.md b/Readme.md
index 753f524..16e89a1 100644
--- a/Readme.md
+++ b/Readme.md
@@ -83,7 +83,7 @@ conda install ccpi-regulariser -c ccpi -c conda-forge
 
 #### Python (conda-build)
 ```
-	export CIL_VERSION=0.10.1
+	export CIL_VERSION=0.10.2
 	conda build Wrappers/Python/conda-recipe --numpy 1.12 --python 3.5 
 	conda install ccpi-regulariser=${CIL_VERSION} --use-local --force
 	cd demos/
diff --git a/Wrappers/Matlab/demos/demoMatlab_denoise.m b/Wrappers/Matlab/demos/demoMatlab_denoise.m
index 2cbdb56..54b8bac 100644
--- a/Wrappers/Matlab/demos/demoMatlab_denoise.m
+++ b/Wrappers/Matlab/demos/demoMatlab_denoise.m
@@ -136,21 +136,20 @@ figure; imshow(u_diff4, [0 1]); title('Diffusion 4thO denoised image (CPU)');
 % 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 = 9;
-PatchWindow = 3;
+SearchingWindow = 7;
+PatchWindow = 2;
 NeighboursNumber = 15; % the number of neibours to include
-h = 0.25; % edge related parameter for NLM
+h = 0.23; % edge related parameter for NLM
 [H_i, H_j, Weights] = PatchSelect(single(u0), SearchingWindow, PatchWindow, NeighboursNumber, h);
 %%
 fprintf('Denoise using Non-local Total Variation (CPU) \n');
-iter_nltv = 3; % number of nltv iterations
+iter_nltv = 2; % number of nltv iterations
 lambda_nltv = 0.085; % 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; imshow(u_nltv, [0 1]); title('Non-local Total Variation denoised image (CPU)');
 %%
-
 %>>>>>>>>>>>>>> MULTI-CHANNEL priors <<<<<<<<<<<<<<< %
 
 fprintf('Denoise using the FGP-dTV model (CPU) \n');
diff --git a/Wrappers/Matlab/mex_compile/compileCPU_mex_Linux.m b/Wrappers/Matlab/mex_compile/compileCPU_mex_Linux.m
index 49b5dfd..72a828e 100644
--- a/Wrappers/Matlab/mex_compile/compileCPU_mex_Linux.m
+++ b/Wrappers/Matlab/mex_compile/compileCPU_mex_Linux.m
@@ -72,6 +72,7 @@ mex NonlocalMarching_Inpaint.c NonlocalMarching_Inpaint_core.c utils.c CFLAGS="\
 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! >>>>>>>');
 
diff --git a/Wrappers/Matlab/mex_compile/compileCPU_mex_WINDOWS.m b/Wrappers/Matlab/mex_compile/compileCPU_mex_WINDOWS.m
index 1b59dc2..6f7541c 100644
--- a/Wrappers/Matlab/mex_compile/compileCPU_mex_WINDOWS.m
+++ b/Wrappers/Matlab/mex_compile/compileCPU_mex_WINDOWS.m
@@ -60,6 +60,12 @@ 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);
@@ -73,9 +79,7 @@ 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);
 
-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 Diffusion_Inpaint_core* NonlocalMarching_Inpaint_core*
-fprintf('%s \n', 'Regularisers successfully compiled!');
+
 %%
 %%% The second approach to compile using TDM-GCC which follows this
 %%% discussion:
@@ -105,15 +109,24 @@ fprintf('%s \n', 'Regularisers successfully compiled!');
 % 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 Diffusion_Inpaint_core* NonlocalMarching_Inpaint_core*
-% fprintf('%s \n', 'Regularisers successfully compiled!');
+
+
+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!');
+
+
 %%
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
diff --git a/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV.c b/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV.c
index dea343c..014c0a0 100644
--- a/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV.c
+++ b/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV.c
@@ -68,7 +68,6 @@ void mexFunction(
     lambda = (float) mxGetScalar(prhs[5]); /* regularisation parameter */
     IterNumb = (int) mxGetScalar(prhs[6]); /* the number of iterations */
  
-    lambda = 1.0f/lambda;       
     dimX = dim_array[0]; dimY = dim_array[1]; dimZ = dim_array[2];   
          
     /*****2D INPUT *****/
@@ -81,9 +80,9 @@ void mexFunction(
     /****************************************************/
     if (number_of_dims == 3) {
         NumNeighb = dim_array2[3];
-        Output = (float*)mxGetPr(plhs[0] = mxCreateNumericArray(3, dim_array, mxSINGLE_CLASS, mxREAL));               
+        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);        
+    Nonlocal_TV_CPU_main(A_orig, Output, H_i, H_j, H_k, Weights, dimX, dimY, dimZ, NumNeighb, lambda, IterNumb);
 }
diff --git a/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV_core.c b/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV_core.c
deleted file mode 100644
index d327dd5..0000000
--- a/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV_core.c
+++ /dev/null
@@ -1,169 +0,0 @@
-/*
- * 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 lambda, int IterNumb)
-{
-
-    long i, j, k;
-    int iter;
-         
-    /*****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, dimX, dimY, NumNeighb, lambda); */ /* NLM - H1 penalty */
-            NLM_TV_2D(Output, A_orig, H_i, H_j, Weights, i, j, (long)(dimX), (long)(dimY), NumNeighb, lambda);   /* 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, lambda); */ /* 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, lambda);   /* 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 lambda)
-{
-	long x, i1, j1, index; 
-	float value = 0.0f, normweight  = 0.0f;
-	
-	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[j*dimX+i] = (lambda*A_orig[j*dimX+i] + value)/(lambda + 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 lambda)
-{
-	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] = (lambda*A_orig[(dimX*dimY*k) + j*dimX+i] + value)/(lambda + 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 lambda)
-{
-	long x, i1, j1, index; 
-	float value = 0.0f, normweight  = 0.0f, NLgrad_magn = 0.0f, NLCoeff;
-	
-	for(x=0; x < NumNeighb; x++) {
-		index =  (dimX*dimY*x) + j*dimX+i;
-		i1 = H_i[index];
-		j1 = H_j[index];
-	        NLgrad_magn += powf((A[j1*dimX+i1] - A[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*x) + j*dimX+i;
-	i1 = H_i[index];
-	j1 = H_j[index];
-        value += A[j1*dimX+i1]*NLCoeff*Weights[index];
-        normweight += Weights[index]*NLCoeff;
-    }   		
-    A[j*dimX+i] = (lambda*A_orig[j*dimX+i] + value)/(lambda + 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 lambda)
-{
-	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] = (lambda*A_orig[(dimX*dimY*k) + j*dimX+i] + value)/(lambda + normweight);
-    return *A;
-}
diff --git a/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV_core.h b/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV_core.h
deleted file mode 100644
index 5b6963e..0000000
--- a/Wrappers/Matlab/mex_compile/regularisers_CPU/Nonlocal_TV_core.h
+++ /dev/null
@@ -1,61 +0,0 @@
-/*
- * 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 lambda, 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 lambda);
-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 lambda);
-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 lambda);
-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 lambda);
-#ifdef __cplusplus
-}
-#endif
diff --git a/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect.c b/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect.c
index fdd9a97..f942539 100644
--- a/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect.c
+++ b/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect.c
@@ -87,6 +87,6 @@ void mexFunction(
         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); 
+    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/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect_core.c b/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect_core.c
deleted file mode 100644
index efc5498..0000000
--- a/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect_core.c
+++ /dev/null
@@ -1,254 +0,0 @@
-/*
- * 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
- */
-
-/**************************************************/
-
-float PatchSelect_CPU_main(float *A, unsigned short *H_i, unsigned short *H_j, unsigned short *H_k, float *Weights, long dimX, long dimY, long dimZ, int SearchWindow, int SimilarWin, int NumNeighb, float h)
-{
-    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) */
-#pragma omp parallel for shared (A, Weights, H_i, H_j) private(i,j)
-        for(i=0; i<dimX; i++) {
-            for(j=0; j<dimY; j++) {
-                Indeces2D(A, H_i, H_j, Weights, (i), (j), (dimX), (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) */
-#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 dimY, long dimX, 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, temp;
-    unsigned short *ind_i, *ind_j, temp_i, temp_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; 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];
-                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;
-            }}}
-    /*sorting loop finished*/
-    
-    /*now select the NumNeighb more prominent weights and store into 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 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/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect_core.h b/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect_core.h
deleted file mode 100644
index 43fce87..0000000
--- a/Wrappers/Matlab/mex_compile/regularisers_CPU/PatchSelect_core.h
+++ /dev/null
@@ -1,62 +0,0 @@
-/*
- * 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, long dimX, long dimY, long dimZ, int SearchWindow, int SimilarWin, int NumNeighb, float h);
-CCPI_EXPORT float Indeces2D(float *Aorig, unsigned short *H_i, unsigned short *H_j, float *Weights, long i, long j, long dimY, long dimX, 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/Wrappers/Python/ccpi/filters/regularisers.py b/Wrappers/Python/ccpi/filters/regularisers.py
index c3c3c7e..bf7e23c 100644
--- a/Wrappers/Python/ccpi/filters/regularisers.py
+++ b/Wrappers/Python/ccpi/filters/regularisers.py
@@ -2,7 +2,7 @@
 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
+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
     gpu_enabled = True
@@ -145,7 +145,7 @@ def DIFF4th(inputData, regularisation_parameter, edge_parameter, iterations,
         raise ValueError('Unknown device {0}. Expecting gpu or cpu'\
                          .format(device))
         
-def PatchSelect_CPU(inputData, searchwindow, patchwindow, neighbours, edge_parameter, device='cpu'):
+def PatchSelect(inputData, searchwindow, patchwindow, neighbours, edge_parameter, device='cpu'):
     if device == 'cpu':
         return PATCHSEL_CPU(inputData,
                      searchwindow,
@@ -159,7 +159,16 @@ def PatchSelect_CPU(inputData, searchwindow, patchwindow, neighbours, edge_param
     	    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':
diff --git a/Wrappers/Python/demos/demo_cpu_regularisers.py b/Wrappers/Python/demos/demo_cpu_regularisers.py
index e99b271..31e4cad 100644
--- a/Wrappers/Python/demos/demo_cpu_regularisers.py
+++ b/Wrappers/Python/demos/demo_cpu_regularisers.py
@@ -13,6 +13,7 @@ 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):
@@ -350,7 +351,7 @@ a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14,
 imgplot = plt.imshow(ndf_cpu, cmap="gray")
 plt.title('{}'.format('CPU results'))
 
-
+#%%
 print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
 print ("___Anisotropic Diffusion 4th Order (2D)____")
 print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
@@ -395,7 +396,71 @@ a.text(0.15, 0.25, txtstr, transform=a.transAxes, fontsize=14,
 imgplot = plt.imshow(diff4_cpu, cmap="gray")
 plt.title('{}'.format('CPU results'))
 
+#%%
+print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+print ("___Nonlocal patches pre-calculation____")
+print ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
+# set parameters
+pars = {'algorithm' : PatchSelect, \
+        'input' : u0,\
+        'searchwindow': 7, \
+        'patchwindow': 2,\
+        'neighbours' : 15 ,\
+        'edge_parameter':0.23}
+
+H_i, H_j, Weights = PatchSelect(pars['input'], 
+              pars['searchwindow'],
+              pars['patchwindow'], 
+              pars['neighbours'],
+              pars['edge_parameter'],'cpu')
+
+#%%
+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': 0.085,\
+        'iterations': 2
+        }
+#%%
+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 ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
@@ -447,7 +512,7 @@ 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 ("%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%")
diff --git a/Wrappers/Python/src/cpu_regularisers.pyx b/Wrappers/Python/src/cpu_regularisers.pyx
index b056bba..e51e6d8 100644
--- a/Wrappers/Python/src/cpu_regularisers.pyx
+++ b/Wrappers/Python/src/cpu_regularisers.pyx
@@ -27,10 +27,9 @@ cdef extern float Diffusion_CPU_main(float *Input, float *Output, float lambdaPa
 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, long dimX, long dimY, long dimZ, int SearchWindow, int SimilarWin, int NumNeighb, float h);
+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);
@@ -450,7 +449,6 @@ def Diff4th_3D(np.ndarray[np.float32_t, ndim=3, mode="c"] inputData,
 
     return outputData
 
-
 #****************************************************************#
 #***************Patch-based weights calculation******************#
 #****************************************************************#
@@ -458,31 +456,85 @@ def PATCHSEL_CPU(inputData, searchwindow, patchwindow, neighbours, edge_paramete
     if inputData.ndim == 2:
         return PatchSel_2D(inputData, searchwindow, patchwindow, neighbours, edge_parameter)
     elif inputData.ndim == 3:
-        return 1
-# PatchSel_3D(inputData, searchwindow, patchwindow, neighbours, edge_parameter)
-def PatchSel_2D(np.ndarray[np.float32_t, ndim=2, mode="c"] inputData, 
+        return PatchSel_3D(inputData, searchwindow, patchwindow, neighbours, edge_parameter)
+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] = inputData.shape[0]
-    dims[1] = inputData.shape[1]
-    dims[2] = neighbours
+    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')
+            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 ')
+            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 ')
+            np.zeros([dims[0], dims[1],dims[2]], dtype='uint16')
 
     # Run patch-based weight selection function
-    PatchSelect_CPU_main(&inputData[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], 1, searchwindow, patchwindow,  neighbours,  edge_parameter)
+    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******************#
diff --git a/run.sh b/run.sh
new file mode 100644
index 0000000..98b792e
--- /dev/null
+++ b/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