-
Notifications
You must be signed in to change notification settings - Fork 0
/
vol_math_anistropic.cpp
315 lines (315 loc) · 12.6 KB
/
vol_math_anistropic.cpp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
//#include "vol_math_RawImage.h"
//#include "vol_math_Raw3D_Independt.h"
//#include "vol_math_anistropic.h"
//
//
//
//PIXTYPE max_min( PIXTYPE val_min )
//{
// throw std::exception("The method or operation is not implemented.");
//}
//
//PIXTYPE Anistropic:: _linear_atXY(const float fx, const float fy, const int z, const int c, const PIXTYPE out_value)
// {
// const int
// x = (int)fx - (fx>=0?0:1), nx = x + 1,
// y = (int)fy - (fy>=0?0:1), ny = y + 1;
// const float
// dx = fx - x,
// dy = fy - y;
// const PIXTYPE
// Icc = (PIXTYPE)atXY(x,y,z,c,out_value), Inc = (PIXTYPE)atXY(nx,y,z,c,out_value),
// Icn = (PIXTYPE)atXY(x,ny,z,c,out_value), Inn = (PIXTYPE)atXY(nx,ny,z,c,out_value);
// return Icc + dx*(Inc-Icc + dy*(Icc+Inn-Icn-Inc)) + dy*(Icn-Icc);
//}
//
//PIXTYPE Anistropic:: _linear_atXYZ(const float fx, const float fy=0, const float fz=0, const int c=0) {
// const float
// nfx = fx<0?0:(fx>_width-1?_width-1:fx),
// nfy = fy<0?0:(fy>_height-1?_height-1:fy),
// nfz = fz<0?0:(fz>_depth-1?_depth-1:fz);
// const unsigned int
// x = (unsigned int)nfx,
// y = (unsigned int)nfy,
// z = (unsigned int)nfz;
// const float
// dx = nfx - x,
// dy = nfy - y,
// dz = nfz - z;
// const unsigned int
// nx = dx>0?x+1:x,
// ny = dy>0?y+1:y,
// nz = dz>0?z+1:z;
// const PIXTYPE
// Iccc = (T)(raw4d)(x,y,z,c), Incc = (T)(raw4d)(nx,y,z,c),
// Icnc = (T)(raw4d)(x,ny,z,c), Innc = (T)(raw4d)(nx,ny,z,c),
// Iccn = (T)(raw4d)(x,y,nz,c), Incn = (T)(raw4d)(nx,y,nz,c),
// Icnn = (T)(raw4d)(x,ny,nz,c), Innn = (T)(raw4d)(nx,ny,nz,c);
// return Iccc +
// dx*(Incc-Iccc +
// dy*(Iccc+Innc-Icnc-Incc +
// dz*(Iccn+Innn+Icnc+Incc-Icnn-Incn-Iccc-Innc)) +
// dz*(Iccc+Incn-Iccn-Incc)) +
// dy*(Icnc-Iccc +
// dz*(Iccc+Icnn-Iccn-Icnc)) +
// dz*(Iccn-Iccc);
// }
//
////void raw2d_3x3( Raw2D I, PIXTYPE T )
////{
//// throw std::exception("The method or operation is not implemented.");
////}
//
// Raw4D& blur_anisotropic( Raw4D & G,
// const float amplitude=60, const float dl=0.8f, const float da=30,
// const float gauss_prec=2, const unsigned int interpolation_type=0,
// const bool is_fast_approx=1)
// {
// Raw4D ret(G);
// // Check arguments and init variables
// /*if (!is_sameXYZ(G) || (G._spectrum!=3 && G._spectrum!=6))
// throw CImgArgumentException(_cimg_instance
// "blur_anisotropic(): Invalid specified diffusion tensor field (%u,%u,%u,%u,%p).",
// cimg_instance,
// G._width,G._height,G._depth,G._spectrum,G._data);*/
//
// //if (is_empty() || amplitude<=0 || dl<0) return *this;
// const bool is_3d = 1;
// PIXTYPE val_min, val_max = max_min(val_min);
// int _width,_height,_depth,_spectrum;
// if (da<=0) { // Iterated oriented Laplacians
//
// Raw4D velocity(_width,_height,_depth);
// for (unsigned int iteration = 0; iteration<(unsigned int)amplitude; ++iteration) {
// PIXTYPE *ptrd = velocity.getdata(), veloc_max = 0;
// if (is_3d) { // 3d version
// RAW_3x3x3(I,PIXTYPE);
// RAW_forC(ret,c) RAW_for3x3x3(ret,x,y,z,c,I,PIXTYPE) {
// const PIXTYPE
// ixx = Incc + Ipcc - 2*Iccc,
// ixy = (Innc + Ippc - Inpc - Ipnc)/4,
// ixz = (Incn + Ipcp - Incp - Ipcn)/4,
// iyy = Icnc + Icpc - 2*Iccc,
// iyz = (Icnn + Icpp - Icnp - Icpn)/4,
// izz = Iccn + Iccp - 2*Iccc,
// veloc = (PIXTYPE)(G.get(x,y,z,0)*ixx + 2*G.get(x,y,z,1)*ixy + 2*G(x,y,z,2)*ixz + G(x,y,z,3)*iyy + 2*G(x,y,z,4)*iyz + G(x,y,z,5)*izz);
// *(ptrd++) = veloc;
// if (veloc>veloc_max) veloc_max = veloc; else if (-veloc>veloc_max) veloc_max = -veloc;
// }
// } else { // 2d version
// Raw2d_3x3(I,PIXTYPE);
// RAW_forZC(ret,z,c) RAW_for3x3(ret,x,y,z,c,I,PIXTYPE) {
// const PIXTYPE
// ixx = Inc + Ipc - 2*Icc,
// ixy = (Inn + Ipp - Inp - Ipn)/4,
// iyy = Icn + Icp - 2*Icc,
// veloc = (PIXTYPE)(G(x,y,0,0)*ixx + 2*G(x,y,0,1)*ixy + G(x,y,0,2)*iyy);
// *(ptrd++) = veloc;
// if (veloc>veloc_max) veloc_max = veloc; else if (-veloc>veloc_max) veloc_max = -veloc;
// }
// }
// if (veloc_max>0) ret+=(velocity*=dl/veloc_max);
// }
// } else { // LIC-based smoothing.
// const unsigned long whd = (unsigned long)_width*_height*_depth;
// const float sqrt2amplitude = (float)std::sqrt(2*amplitude);
// const int dx1 = G.getXsize() - 1, dy1 = G.getYsize() - 1, dz1 = G.getZsize() - 1;
// Raw4D res(_width,_height,_depth,_spectrum,0), W(_width,_height,_depth,is_3d?4:3), val(_spectrum);
// int N = 0;
// if (is_3d) { // 3d version
// for (float phi = (180%(int)da)/2.0f; phi<=180; phi+=da) {
// const float phir = (float)(phi*PI/180), datmp = (float)(da/std::cos(phir)), da2 = datmp<1?360.0f:datmp;
// for (float theta = 0; theta<360; (theta+=da2),++N) {
// const float
// thetar = (float)(theta*PI/180),
// vx = (float)(std::cos(thetar)*std::cos(phir)),
// vy = (float)(std::sin(thetar)*std::cos(phir)),
// vz = (float)std::sin(phir);
// const PIXTYPE
// *pa = G.get(0,0,0), *pb = G.data(0,0,0,1), *pc = G.data(0,0,0,2),
// *pd = G.data(0,0,0,3), *pe = G.data(0,0,0,4), *pf = G.data(0,0,0,5);
// PIXTYPE *pd0 = W.data(0,0,0,0), *pd1 = W.data(0,0,0,1), *pd2 = W.data(0,0,0,2), *pd3 = W.data(0,0,0,3);
// RAW_forXYZ(ret,xg,yg,zg) {
// const PIXTYPE a = *(pa++), b = *(pb++), c = *(pc++), d = *(pd++), e = *(pe++), f = *(pf++);
// const float
// u = (float)(a*vx + b*vy + c*vz),
// v = (float)(b*vx + d*vy + e*vz),
// w = (float)(c*vx + e*vy + f*vz),
// n = (float)std::sqrt(1e-5+u*u+v*v+w*w),
// dln = dl/n;
// *(pd0++) = (PIXTYPE)(u*dln);
// *(pd1++) = (PIXTYPE)(v*dln);
// *(pd2++) = (PIXTYPE)(w*dln);
// *(pd3++) = (PIXTYPE)n;
// }
//
// PIXTYPE *ptrd = res.getdata();
// RAW_forXYZ(ret,x,y,z) {
// val.put(0);
// const float
// n = (float)W(x,y,z,3),
// fsigma = (float)(n*sqrt2amplitude),
// fsigma2 = 2*fsigma*fsigma,
// length = gauss_prec*fsigma;
// float
// S = 0,
// X = (float)x,
// Y = (float)y,
// Z = (float)z;
// switch (interpolation_type) {
// case 0 : { // Nearest neighbor
// for (float l = 0; l<length && X>=0 && X<=dx1 && Y>=0 && Y<=dy1 && Z>=0 && Z<=dz1; l+=dl) {
// const int
// cx = (int)(X+0.5f),
// cy = (int)(Y+0.5f),
// cz = (int)(Z+0.5f);
// const float
// u = (float)W(cx,cy,cz,0),
// v = (float)W(cx,cy,cz,1),
// w = (float)W(cx,cy,cz,2);
// if (is_fast_approx) { RAW_forC(*this,c) val[c]+=(PIXTYPE)(ret)(cx,cy,cz,c); ++S; }
// else {
// const float coef = (float)std::exp(-l*l/fsigma2);
// RAW_forC(ret,c) val[c]+=(PIXTYPE)(coef*(ret)(cx,cy,cz,c));
// S+=coef;
// }
// X+=u; Y+=v; Z+=w;
// }
// } break;
// case 1 : { // Linear interpolation
// for (float l = 0; l<length && X>=0 && X<=dx1 && Y>=0 && Y<=dy1 && Z>=0 && Z<=dz1; l+=dl) {
// const float
// u = (float)(W._linear_atXYZ(X,Y,Z,0)),
// v = (float)(W._linear_atXYZ(X,Y,Z,1)),
// w = (float)(W._linear_atXYZ(X,Y,Z,2));
// if (is_fast_approx) {RAW_forC(*this,c) val[c]+=(PIXTYPE)_linear_atXYZ(X,Y,Z,c); ++S; }
// else {
// const float coef = (float)std::exp(-l*l/fsigma2);
// RAW_forC(*this,c) val[c]+=(PIXTYPE)(coef*_linear_atXYZ(X,Y,Z,c));
// S+=coef;
// }
// X+=u; Y+=v; Z+=w;
// }
// } break;
// default : { // 2nd order Runge Kutta
// for (float l = 0; l<length && X>=0 && X<=dx1 && Y>=0 && Y<=dy1 && Z>=0 && Z<=dz1; l+=dl) {
// const float
// u0 = (float)(0.5f*W._linear_atXYZ(X,Y,Z,0)),
// v0 = (float)(0.5f*W._linear_atXYZ(X,Y,Z,1)),
// w0 = (float)(0.5f*W._linear_atXYZ(X,Y,Z,2)),
// u = (float)(W._linear_atXYZ(X+u0,Y+v0,Z+w0,0)),
// v = (float)(W._linear_atXYZ(X+u0,Y+v0,Z+w0,1)),
// w = (float)(W._linear_atXYZ(X+u0,Y+v0,Z+w0,2));
// if (is_fast_approx) { RAW_forC(*this,c) val[c]+=(PIXTYPE)_linear_atXYZ(X,Y,Z,c); ++S; }
// else {
// const float coef = (float)std::exp(-l*l/fsigma2);
// RAW_forC(*this,c) val[c]+=(PIXTYPE)(coef*_linear_atXYZ(X,Y,Z,c));
// S+=coef;
// }
// X+=u; Y+=v; Z+=w;
// }
// } break;
// }
// PIXTYPE *_ptrd = ptrd++;
// if (S>0) RAW_forC(res,c) { *_ptrd+=val[c]/S; _ptrd+=whd; }
// else RAW_forC(res,c) { *_ptrd+=(PIXTYPE)((ret)(x,y,z,c)); _ptrd+=whd; }
// }
// }
// }
// } else { // 2d LIC algorithm
// for (float theta = (360%(int)da)/2.0f; theta<360; (theta+=da),++N) {
// const float thetar = (float)(theta*PI/180), vx = (float)(std::cos(thetar)), vy = (float)(std::sin(thetar));
// const PIXTYPE *pa = G.data(0,0,0,0), *pb = G.data(0,0,0,1), *pc = G.data(0,0,0,2);
// PIXTYPE *pd0 = W.data(0,0,0,0), *pd1 = W.data(0,0,0,1), *pd2 = W.data(0,0,0,2);
// RAW_forXY(G,xg,yg) {
// const PIXTYPE a = *(pa++), b = *(pb++), c = *(pc++);
// const float
// u = (float)(a*vx + b*vy),
// v = (float)(b*vx + c*vy),
// n = (float)std::sqrt(1e-5+u*u+v*v),
// dln = dl/n;
// *(pd0++) = (PIXTYPE)(u*dln);
// *(pd1++) = (PIXTYPE)(v*dln);
// *(pd2++) = (PIXTYPE)n;
// }
// PIXTYPE *ptrd = res.getdata();
// RAW_forXY(ret,x,y) {
// val.fill(0);
// const float
// n = (float)W(x,y,0,2),
// fsigma = (float)(n*sqrt2amplitude),
// fsigma2 = 2*fsigma*fsigma,
// length = gauss_prec*fsigma;
// float
// S = 0,
// X = (float)x,
// Y = (float)y;
// switch (interpolation_type) {
// case 0 : { // Nearest-neighbor
// for (float l = 0; l<length && X>=0 && X<=dx1 && Y>=0 && Y<=dy1; l+=dl) {
// const int
// cx = (int)(X+0.5f),
// cy = (int)(Y+0.5f);
// const float
// u = (float)W(cx,cy,0,0),
// v = (float)W(cx,cy,0,1);
// if (is_fast_approx) { RAW_forC(*this,c) val[c]+=(PIXTYPE)(ret)(cx,cy,0,c); ++S; }
// else {
// const float coef = (float)std::exp(-l*l/fsigma2);
// RAW_forC(*this,c) val[c]+=(PIXTYPE)(coef*(ret)(cx,cy,0,c));
// S+=coef;
// }
// X+=u; Y+=v;
// }
// } break;
// case 1 : { // Linear interpolation
// for (float l = 0; l<length && X>=0 && X<=dx1 && Y>=0 && Y<=dy1; l+=dl) {
// const float
// u = (float)(W._linear_atXY(X,Y,0,0)),
// v = (float)(W._linear_atXY(X,Y,0,1));
// if (is_fast_approx) { RAW_forC(*this,c) val[c]+=(PIXTYPE)_linear_atXY(X,Y,0,c); ++S; }
// else {
// const float coef = (float)std::exp(-l*l/fsigma2);
// RAW_forC(*this,c) val[c]+=(PIXTYPE)(coef*_linear_atXY(X,Y,0,c));
// S+=coef;
// }
// X+=u; Y+=v;
// }
// } break;
// default : { // 2nd-order Runge-kutta interpolation
// for (float l = 0; l<length && X>=0 && X<=dx1 && Y>=0 && Y<=dy1; l+=dl) {
// const float
// u0 = (float)(0.5f*W._linear_atXY(X,Y,0,0)),
// v0 = (float)(0.5f*W._linear_atXY(X,Y,0,1)),
// u = (float)(W._linear_atXY(X+u0,Y+v0,0,0)),
// v = (float)(W._linear_atXY(X+u0,Y+v0,0,1));
// if (is_fast_approx) { RAW_forC(*this,c) val[c]+=(PIXTYPE)_linear_atXY(X,Y,0,c); ++S; }
// else {
// const float coef = (float)std::exp(-l*l/fsigma2);
// RAW_forC(*this,c) val[c]+=(PIXTYPE)(coef*_linear_atXY(X,Y,0,c));
// S+=coef;
// }
// X+=u; Y+=v;
// }
// }
// }
// PIXTYPE *_ptrd = ptrd++;
// if (S>0) RAW_forC(res,c) { *_ptrd+=val[c]/S; _ptrd+=whd; }
// else RAW_forC(res,c) { *_ptrd+=(PIXTYPE)((ret)(x,y,0,c)); _ptrd+=whd; }
// }
// }
// }
// const PIXTYPE *ptrs = res.getdata();
// RAW_for(G,ptrd,PIXTYPE) { const PIXTYPE val = *(ptrs++)/N; *ptrd = val<val_min?val_min:(val>val_max?val_max:(PIXTYPE)val); }
// }
// return G;
// }
//
// //! Blur image anisotropically, directed by a field of diffusion tensors \newinstance.
// template<typename PIXTYPE>
// CImg<PIXTYPE> get_blur_anisotropic(const CImg<PIXTYPE>& G,
// const float amplitude=60, const float dl=0.8f, const float da=30,
// const float gauss_prec=2, const unsigned int interpolation_type=0,
// const bool is_fast_approx=true) const {
// return (+*this).blur_anisotropic(G,amplitude,dl,da,gauss_prec,interpolation_type,is_fast_approx);
// }