• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /*M///////////////////////////////////////////////////////////////////////////////////////
2 //
3 //  IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
4 //
5 //  By downloading, copying, installing or using the software you agree to this license.
6 //  If you do not agree to this license, do not download, install,
7 //  copy or use the software.
8 //
9 //
10 //                        Intel License Agreement
11 //                For Open Source Computer Vision Library
12 //
13 // Copyright (C) 2000, Intel Corporation, all rights reserved.
14 // Third party copyrights are property of their respective owners.
15 //
16 // Redistribution and use in source and binary forms, with or without modification,
17 // are permitted provided that the following conditions are met:
18 //
19 //   * Redistribution's of source code must retain the above copyright notice,
20 //     this list of conditions and the following disclaimer.
21 //
22 //   * Redistribution's in binary form must reproduce the above copyright notice,
23 //     this list of conditions and the following disclaimer in the documentation
24 //     and/or other materials provided with the distribution.
25 //
26 //   * The name of Intel Corporation may not be used to endorse or promote products
27 //     derived from this software without specific prior written permission.
28 //
29 // This software is provided by the copyright holders and contributors "as is" and
30 // any express or implied warranties, including, but not limited to, the implied
31 // warranties of merchantability and fitness for a particular purpose are disclaimed.
32 // In no event shall the Intel Corporation or contributors be liable for any direct,
33 // indirect, incidental, special, exemplary, or consequential damages
34 // (including, but not limited to, procurement of substitute goods or services;
35 // loss of use, data, or profits; or business interruption) however caused
36 // and on any theory of liability, whether in contract, strict liability,
37 // or tort (including negligence or otherwise) arising in any way out of
38 // the use of this software, even if advised of the possibility of such damage.
39 //
40 //M*/
41 
42 /****************************************************************************************\
43 
44       Calculation of a texture descriptors from GLCM (Grey Level Co-occurrence Matrix'es)
45       The code was submitted by Daniel Eaton [danieljameseaton@yahoo.com]
46 
47 \****************************************************************************************/
48 
49 #include "_cvaux.h"
50 
51 #include <math.h>
52 #include <assert.h>
53 
54 #define CV_MAX_NUM_GREY_LEVELS_8U  256
55 
56 struct CvGLCM
57 {
58     int matrixSideLength;
59     int numMatrices;
60     double*** matrices;
61 
62     int  numLookupTableElements;
63     int  forwardLookupTable[CV_MAX_NUM_GREY_LEVELS_8U];
64     int  reverseLookupTable[CV_MAX_NUM_GREY_LEVELS_8U];
65 
66     double** descriptors;
67     int numDescriptors;
68     int descriptorOptimizationType;
69     int optimizationType;
70 };
71 
72 
73 static void icvCreateGLCM_LookupTable_8u_C1R( const uchar* srcImageData, int srcImageStep,
74                                              CvSize srcImageSize, CvGLCM* destGLCM,
75                                              int* steps, int numSteps, int* memorySteps );
76 
77 static void
78 icvCreateGLCMDescriptors_AllowDoubleNest( CvGLCM* destGLCM, int matrixIndex );
79 
80 
81 CV_IMPL CvGLCM*
cvCreateGLCM(const IplImage * srcImage,int stepMagnitude,const int * srcStepDirections,int numStepDirections,int optimizationType)82 cvCreateGLCM( const IplImage* srcImage,
83               int stepMagnitude,
84               const int* srcStepDirections,/* should be static array..
85                                           or if not the user should handle de-allocation */
86               int numStepDirections,
87               int optimizationType )
88 {
89     static const int defaultStepDirections[] = { 0,1, -1,1, -1,0, -1,-1 };
90 
91     int* memorySteps = 0;
92     CvGLCM* newGLCM = 0;
93     int* stepDirections = 0;
94 
95     CV_FUNCNAME( "cvCreateGLCM" );
96 
97     __BEGIN__;
98 
99     uchar* srcImageData = 0;
100     CvSize srcImageSize;
101     int srcImageStep;
102     int stepLoop;
103     const int maxNumGreyLevels8u = CV_MAX_NUM_GREY_LEVELS_8U;
104 
105     if( !srcImage )
106         CV_ERROR( CV_StsNullPtr, "" );
107 
108     if( srcImage->nChannels != 1 )
109         CV_ERROR( CV_BadNumChannels, "Number of channels must be 1");
110 
111     if( srcImage->depth != IPL_DEPTH_8U )
112         CV_ERROR( CV_BadDepth, "Depth must be equal IPL_DEPTH_8U");
113 
114     // no Directions provided, use the default ones - 0 deg, 45, 90, 135
115     if( !srcStepDirections )
116     {
117         srcStepDirections = defaultStepDirections;
118     }
119 
120     CV_CALL( stepDirections = (int*)cvAlloc( numStepDirections*2*sizeof(stepDirections[0])));
121     memcpy( stepDirections, srcStepDirections, numStepDirections*2*sizeof(stepDirections[0]));
122 
123     cvGetImageRawData( srcImage, &srcImageData, &srcImageStep, &srcImageSize );
124 
125     // roll together Directions and magnitudes together with knowledge of image (step)
126     CV_CALL( memorySteps = (int*)cvAlloc( numStepDirections*sizeof(memorySteps[0])));
127 
128     for( stepLoop = 0; stepLoop < numStepDirections; stepLoop++ )
129     {
130         stepDirections[stepLoop*2 + 0] *= stepMagnitude;
131         stepDirections[stepLoop*2 + 1] *= stepMagnitude;
132 
133         memorySteps[stepLoop] = stepDirections[stepLoop*2 + 0]*srcImageStep +
134                                 stepDirections[stepLoop*2 + 1];
135     }
136 
137     CV_CALL( newGLCM = (CvGLCM*)cvAlloc(sizeof(newGLCM)));
138     memset( newGLCM, 0, sizeof(newGLCM) );
139 
140     newGLCM->matrices = 0;
141     newGLCM->numMatrices = numStepDirections;
142     newGLCM->optimizationType = optimizationType;
143 
144     if( optimizationType <= CV_GLCM_OPTIMIZATION_LUT )
145     {
146         int lookupTableLoop, imageColLoop, imageRowLoop, lineOffset = 0;
147 
148         // if optimization type is set to lut, then make one for the image
149         if( optimizationType == CV_GLCM_OPTIMIZATION_LUT )
150         {
151             for( imageRowLoop = 0; imageRowLoop < srcImageSize.height;
152                                    imageRowLoop++, lineOffset += srcImageStep )
153             {
154                 for( imageColLoop = 0; imageColLoop < srcImageSize.width; imageColLoop++ )
155                 {
156                     newGLCM->forwardLookupTable[srcImageData[lineOffset+imageColLoop]]=1;
157                 }
158             }
159 
160             newGLCM->numLookupTableElements = 0;
161 
162             for( lookupTableLoop = 0; lookupTableLoop < maxNumGreyLevels8u; lookupTableLoop++ )
163             {
164                 if( newGLCM->forwardLookupTable[ lookupTableLoop ] != 0 )
165                 {
166                     newGLCM->forwardLookupTable[ lookupTableLoop ] =
167                         newGLCM->numLookupTableElements;
168                     newGLCM->reverseLookupTable[ newGLCM->numLookupTableElements ] =
169                         lookupTableLoop;
170 
171                     newGLCM->numLookupTableElements++;
172                 }
173             }
174         }
175         // otherwise make a "LUT" which contains all the gray-levels (for code-reuse)
176         else if( optimizationType == CV_GLCM_OPTIMIZATION_NONE )
177         {
178             for( lookupTableLoop = 0; lookupTableLoop <maxNumGreyLevels8u; lookupTableLoop++ )
179             {
180                 newGLCM->forwardLookupTable[ lookupTableLoop ] = lookupTableLoop;
181                 newGLCM->reverseLookupTable[ lookupTableLoop ] = lookupTableLoop;
182             }
183             newGLCM->numLookupTableElements = maxNumGreyLevels8u;
184         }
185 
186         newGLCM->matrixSideLength = newGLCM->numLookupTableElements;
187         icvCreateGLCM_LookupTable_8u_C1R( srcImageData, srcImageStep, srcImageSize,
188                                           newGLCM, stepDirections,
189                                           numStepDirections, memorySteps );
190     }
191     else if( optimizationType == CV_GLCM_OPTIMIZATION_HISTOGRAM )
192     {
193         CV_ERROR( CV_StsBadFlag, "Histogram-based method is not implemented" );
194 
195     /*  newGLCM->numMatrices *= 2;
196         newGLCM->matrixSideLength = maxNumGreyLevels8u*2;
197 
198         icvCreateGLCM_Histogram_8uC1R( srcImageStep, srcImageSize, srcImageData,
199                                        newGLCM, numStepDirections,
200                                        stepDirections, memorySteps );
201     */
202     }
203 
204     __END__;
205 
206     cvFree( &memorySteps );
207     cvFree( &stepDirections );
208 
209     if( cvGetErrStatus() < 0 )
210     {
211         cvFree( &newGLCM );
212     }
213 
214     return newGLCM;
215 }
216 
217 
218 CV_IMPL void
cvReleaseGLCM(CvGLCM ** GLCM,int flag)219 cvReleaseGLCM( CvGLCM** GLCM, int flag )
220 {
221     CV_FUNCNAME( "cvReleaseGLCM" );
222 
223     __BEGIN__;
224 
225     int matrixLoop;
226 
227     if( !GLCM )
228         CV_ERROR( CV_StsNullPtr, "" );
229 
230     if( *GLCM )
231         EXIT; // repeated deallocation: just skip it.
232 
233     if( (flag == CV_GLCM_GLCM || flag == CV_GLCM_ALL) && (*GLCM)->matrices )
234     {
235         for( matrixLoop = 0; matrixLoop < (*GLCM)->numMatrices; matrixLoop++ )
236         {
237             if( (*GLCM)->matrices[ matrixLoop ] )
238             {
239                 cvFree( (*GLCM)->matrices[matrixLoop] );
240                 cvFree( (*GLCM)->matrices + matrixLoop );
241             }
242         }
243 
244         cvFree( &((*GLCM)->matrices) );
245     }
246 
247     if( (flag == CV_GLCM_DESC || flag == CV_GLCM_ALL) && (*GLCM)->descriptors )
248     {
249         for( matrixLoop = 0; matrixLoop < (*GLCM)->numMatrices; matrixLoop++ )
250         {
251             cvFree( (*GLCM)->descriptors + matrixLoop );
252         }
253         cvFree( &((*GLCM)->descriptors) );
254     }
255 
256     if( flag == CV_GLCM_ALL )
257     {
258         cvFree( GLCM );
259     }
260 
261     __END__;
262 }
263 
264 
265 static void
icvCreateGLCM_LookupTable_8u_C1R(const uchar * srcImageData,int srcImageStep,CvSize srcImageSize,CvGLCM * destGLCM,int * steps,int numSteps,int * memorySteps)266 icvCreateGLCM_LookupTable_8u_C1R( const uchar* srcImageData,
267                                   int srcImageStep,
268                                   CvSize srcImageSize,
269                                   CvGLCM* destGLCM,
270                                   int* steps,
271                                   int numSteps,
272                                   int* memorySteps )
273 {
274     int* stepIncrementsCounter = 0;
275 
276     CV_FUNCNAME( "icvCreateGLCM_LookupTable_8u_C1R" );
277 
278     __BEGIN__;
279 
280     int matrixSideLength = destGLCM->matrixSideLength;
281     int stepLoop, sideLoop1, sideLoop2;
282     int colLoop, rowLoop, lineOffset = 0;
283     double*** matrices = 0;
284 
285     // allocate memory to the matrices
286     CV_CALL( destGLCM->matrices = (double***)cvAlloc( sizeof(matrices[0])*numSteps ));
287     matrices = destGLCM->matrices;
288 
289     for( stepLoop=0; stepLoop<numSteps; stepLoop++ )
290     {
291         CV_CALL( matrices[stepLoop] = (double**)cvAlloc( sizeof(matrices[0])*matrixSideLength ));
292         CV_CALL( matrices[stepLoop][0] = (double*)cvAlloc( sizeof(matrices[0][0])*
293                                                   matrixSideLength*matrixSideLength ));
294 
295         memset( matrices[stepLoop][0], 0, matrixSideLength*matrixSideLength*
296                                           sizeof(matrices[0][0]) );
297 
298         for( sideLoop1 = 1; sideLoop1 < matrixSideLength; sideLoop1++ )
299         {
300             matrices[stepLoop][sideLoop1] = matrices[stepLoop][sideLoop1-1] + matrixSideLength;
301         }
302     }
303 
304     CV_CALL( stepIncrementsCounter = (int*)cvAlloc( numSteps*sizeof(stepIncrementsCounter[0])));
305     memset( stepIncrementsCounter, 0, numSteps*sizeof(stepIncrementsCounter[0]) );
306 
307     // generate GLCM for each step
308     for( rowLoop=0; rowLoop<srcImageSize.height; rowLoop++, lineOffset+=srcImageStep )
309     {
310         for( colLoop=0; colLoop<srcImageSize.width; colLoop++ )
311         {
312             int pixelValue1 = destGLCM->forwardLookupTable[srcImageData[lineOffset + colLoop]];
313 
314             for( stepLoop=0; stepLoop<numSteps; stepLoop++ )
315             {
316                 int col2, row2;
317                 row2 = rowLoop + steps[stepLoop*2 + 0];
318                 col2 = colLoop + steps[stepLoop*2 + 1];
319 
320                 if( col2>=0 && row2>=0 && col2<srcImageSize.width && row2<srcImageSize.height )
321                 {
322                     int memoryStep = memorySteps[ stepLoop ];
323                     int pixelValue2 = destGLCM->forwardLookupTable[ srcImageData[ lineOffset + colLoop + memoryStep ] ];
324 
325                     // maintain symmetry
326                     matrices[stepLoop][pixelValue1][pixelValue2] ++;
327                     matrices[stepLoop][pixelValue2][pixelValue1] ++;
328 
329                     // incremenet counter of total number of increments
330                     stepIncrementsCounter[stepLoop] += 2;
331                 }
332             }
333         }
334     }
335 
336     // normalize matrices. each element is a probability of gray value i,j adjacency in direction/magnitude k
337     for( sideLoop1=0; sideLoop1<matrixSideLength; sideLoop1++ )
338     {
339         for( sideLoop2=0; sideLoop2<matrixSideLength; sideLoop2++ )
340         {
341             for( stepLoop=0; stepLoop<numSteps; stepLoop++ )
342             {
343                 matrices[stepLoop][sideLoop1][sideLoop2] /= double(stepIncrementsCounter[stepLoop]);
344             }
345         }
346     }
347 
348     destGLCM->matrices = matrices;
349 
350     __END__;
351 
352     cvFree( &stepIncrementsCounter );
353 
354     if( cvGetErrStatus() < 0 )
355         cvReleaseGLCM( &destGLCM, CV_GLCM_GLCM );
356 }
357 
358 
359 CV_IMPL void
cvCreateGLCMDescriptors(CvGLCM * destGLCM,int descriptorOptimizationType)360 cvCreateGLCMDescriptors( CvGLCM* destGLCM, int descriptorOptimizationType )
361 {
362     CV_FUNCNAME( "cvCreateGLCMDescriptors" );
363 
364     __BEGIN__;
365 
366     int matrixLoop;
367 
368     if( !destGLCM )
369         CV_ERROR( CV_StsNullPtr, "" );
370 
371     if( !(destGLCM->matrices) )
372         CV_ERROR( CV_StsNullPtr, "Matrices are not allocated" );
373 
374     CV_CALL( cvReleaseGLCM( &destGLCM, CV_GLCM_DESC ));
375 
376     if( destGLCM->optimizationType != CV_GLCM_OPTIMIZATION_HISTOGRAM )
377     {
378         destGLCM->descriptorOptimizationType = destGLCM->numDescriptors = descriptorOptimizationType;
379     }
380     else
381     {
382         CV_ERROR( CV_StsBadFlag, "Histogram-based method is not implemented" );
383 //      destGLCM->descriptorOptimizationType = destGLCM->numDescriptors = CV_GLCMDESC_OPTIMIZATION_HISTOGRAM;
384     }
385 
386     CV_CALL( destGLCM->descriptors = (double**)
387             cvAlloc( destGLCM->numMatrices*sizeof(destGLCM->descriptors[0])));
388 
389     for( matrixLoop = 0; matrixLoop < destGLCM->numMatrices; matrixLoop ++ )
390     {
391         CV_CALL( destGLCM->descriptors[ matrixLoop ] =
392                 (double*)cvAlloc( destGLCM->numDescriptors*sizeof(destGLCM->descriptors[0][0])));
393         memset( destGLCM->descriptors[matrixLoop], 0, destGLCM->numDescriptors*sizeof(double) );
394 
395         switch( destGLCM->descriptorOptimizationType )
396         {
397             case CV_GLCMDESC_OPTIMIZATION_ALLOWDOUBLENEST:
398                 icvCreateGLCMDescriptors_AllowDoubleNest( destGLCM, matrixLoop );
399                 break;
400             default:
401                 CV_ERROR( CV_StsBadFlag,
402                 "descriptorOptimizationType different from CV_GLCMDESC_OPTIMIZATION_ALLOWDOUBLENEST\n"
403                 "is not supported" );
404             /*
405             case CV_GLCMDESC_OPTIMIZATION_ALLOWTRIPLENEST:
406                 icvCreateGLCMDescriptors_AllowTripleNest( destGLCM, matrixLoop );
407                 break;
408             case CV_GLCMDESC_OPTIMIZATION_HISTOGRAM:
409                 if(matrixLoop < destGLCM->numMatrices>>1)
410                     icvCreateGLCMDescriptors_Histogram( destGLCM, matrixLoop);
411                     break;
412             */
413         }
414     }
415 
416     __END__;
417 
418     if( cvGetErrStatus() < 0 )
419         cvReleaseGLCM( &destGLCM, CV_GLCM_DESC );
420 }
421 
422 
423 static void
icvCreateGLCMDescriptors_AllowDoubleNest(CvGLCM * destGLCM,int matrixIndex)424 icvCreateGLCMDescriptors_AllowDoubleNest( CvGLCM* destGLCM, int matrixIndex )
425 {
426     int sideLoop1, sideLoop2;
427     int matrixSideLength = destGLCM->matrixSideLength;
428 
429     double** matrix = destGLCM->matrices[ matrixIndex ];
430     double* descriptors = destGLCM->descriptors[ matrixIndex ];
431 
432     double* marginalProbability =
433         (double*)cvAlloc( matrixSideLength * sizeof(marginalProbability[0]));
434     memset( marginalProbability, 0, matrixSideLength * sizeof(double) );
435 
436     double maximumProbability = 0;
437     double marginalProbabilityEntropy = 0;
438     double correlationMean = 0, correlationStdDeviation = 0, correlationProductTerm = 0;
439 
440     for( sideLoop1=0; sideLoop1<matrixSideLength; sideLoop1++ )
441     {
442         int actualSideLoop1 = destGLCM->reverseLookupTable[ sideLoop1 ];
443 
444         for( sideLoop2=0; sideLoop2<matrixSideLength; sideLoop2++ )
445         {
446             double entryValue = matrix[ sideLoop1 ][ sideLoop2 ];
447 
448             int actualSideLoop2 = destGLCM->reverseLookupTable[ sideLoop2 ];
449             int sideLoopDifference = actualSideLoop1 - actualSideLoop2;
450             int sideLoopDifferenceSquared = sideLoopDifference*sideLoopDifference;
451 
452             marginalProbability[ sideLoop1 ] += entryValue;
453             correlationMean += actualSideLoop1*entryValue;
454 
455             maximumProbability = MAX( maximumProbability, entryValue );
456 
457             if( actualSideLoop2 > actualSideLoop1 )
458             {
459                 descriptors[ CV_GLCMDESC_CONTRAST ] += sideLoopDifferenceSquared * entryValue;
460             }
461 
462             descriptors[ CV_GLCMDESC_HOMOGENITY ] += entryValue / ( 1.0 + sideLoopDifferenceSquared );
463 
464             if( entryValue > 0 )
465             {
466                 descriptors[ CV_GLCMDESC_ENTROPY ] += entryValue * log( entryValue );
467             }
468 
469             descriptors[ CV_GLCMDESC_ENERGY ] += entryValue*entryValue;
470         }
471 
472         if( marginalProbability>0 )
473             marginalProbabilityEntropy += marginalProbability[ actualSideLoop1 ]*log(marginalProbability[ actualSideLoop1 ]);
474     }
475 
476     marginalProbabilityEntropy = -marginalProbabilityEntropy;
477 
478     descriptors[ CV_GLCMDESC_CONTRAST ] += descriptors[ CV_GLCMDESC_CONTRAST ];
479     descriptors[ CV_GLCMDESC_ENTROPY ] = -descriptors[ CV_GLCMDESC_ENTROPY ];
480     descriptors[ CV_GLCMDESC_MAXIMUMPROBABILITY ] = maximumProbability;
481 
482     double HXY = 0, HXY1 = 0, HXY2 = 0;
483 
484     HXY = descriptors[ CV_GLCMDESC_ENTROPY ];
485 
486     for( sideLoop1=0; sideLoop1<matrixSideLength; sideLoop1++ )
487     {
488         double sideEntryValueSum = 0;
489         int actualSideLoop1 = destGLCM->reverseLookupTable[ sideLoop1 ];
490 
491         for( sideLoop2=0; sideLoop2<matrixSideLength; sideLoop2++ )
492         {
493             double entryValue = matrix[ sideLoop1 ][ sideLoop2 ];
494 
495             sideEntryValueSum += entryValue;
496 
497             int actualSideLoop2 = destGLCM->reverseLookupTable[ sideLoop2 ];
498 
499             correlationProductTerm += (actualSideLoop1 - correlationMean) * (actualSideLoop2 - correlationMean) * entryValue;
500 
501             double clusterTerm = actualSideLoop1 + actualSideLoop2 - correlationMean - correlationMean;
502 
503             descriptors[ CV_GLCMDESC_CLUSTERTENDENCY ] += clusterTerm * clusterTerm * entryValue;
504             descriptors[ CV_GLCMDESC_CLUSTERSHADE ] += clusterTerm * clusterTerm * clusterTerm * entryValue;
505 
506             double HXYValue = marginalProbability[ actualSideLoop1 ] * marginalProbability[ actualSideLoop2 ];
507             if( HXYValue>0 )
508             {
509                 double HXYValueLog = log( HXYValue );
510                 HXY1 += entryValue * HXYValueLog;
511                 HXY2 += HXYValue * HXYValueLog;
512             }
513         }
514 
515         correlationStdDeviation += (actualSideLoop1-correlationMean) * (actualSideLoop1-correlationMean) * sideEntryValueSum;
516     }
517 
518     HXY1 =- HXY1;
519     HXY2 =- HXY2;
520 
521     descriptors[ CV_GLCMDESC_CORRELATIONINFO1 ] = ( HXY - HXY1 ) / ( correlationMean );
522     descriptors[ CV_GLCMDESC_CORRELATIONINFO2 ] = sqrt( 1.0 - exp( -2.0 * (HXY2 - HXY ) ) );
523 
524     correlationStdDeviation = sqrt( correlationStdDeviation );
525 
526     descriptors[ CV_GLCMDESC_CORRELATION ] = correlationProductTerm / (correlationStdDeviation*correlationStdDeviation );
527 
528     delete [] marginalProbability;
529 }
530 
531 
cvGetGLCMDescriptor(CvGLCM * GLCM,int step,int descriptor)532 CV_IMPL double cvGetGLCMDescriptor( CvGLCM* GLCM, int step, int descriptor )
533 {
534     double value = DBL_MAX;
535 
536     CV_FUNCNAME( "cvGetGLCMDescriptor" );
537 
538     __BEGIN__;
539 
540     if( !GLCM )
541         CV_ERROR( CV_StsNullPtr, "" );
542 
543     if( !(GLCM->descriptors) )
544         CV_ERROR( CV_StsNullPtr, "" );
545 
546     if( (unsigned)step >= (unsigned)(GLCM->numMatrices))
547         CV_ERROR( CV_StsOutOfRange, "step is not in 0 .. GLCM->numMatrices - 1" );
548 
549     if( (unsigned)descriptor >= (unsigned)(GLCM->numDescriptors))
550         CV_ERROR( CV_StsOutOfRange, "descriptor is not in 0 .. GLCM->numDescriptors - 1" );
551 
552     value = GLCM->descriptors[step][descriptor];
553 
554     __END__;
555 
556     return value;
557 }
558 
559 
560 CV_IMPL void
cvGetGLCMDescriptorStatistics(CvGLCM * GLCM,int descriptor,double * _average,double * _standardDeviation)561 cvGetGLCMDescriptorStatistics( CvGLCM* GLCM, int descriptor,
562                                double* _average, double* _standardDeviation )
563 {
564     CV_FUNCNAME( "cvGetGLCMDescriptorStatistics" );
565 
566     if( _average )
567         *_average = DBL_MAX;
568 
569     if( _standardDeviation )
570         *_standardDeviation = DBL_MAX;
571 
572     __BEGIN__;
573 
574     int matrixLoop, numMatrices;
575     double average = 0, squareSum = 0;
576 
577     if( !GLCM )
578         CV_ERROR( CV_StsNullPtr, "" );
579 
580     if( !(GLCM->descriptors))
581         CV_ERROR( CV_StsNullPtr, "Descriptors are not calculated" );
582 
583     if( (unsigned)descriptor >= (unsigned)(GLCM->numDescriptors) )
584         CV_ERROR( CV_StsOutOfRange, "Descriptor index is out of range" );
585 
586     numMatrices = GLCM->numMatrices;
587 
588     for( matrixLoop = 0; matrixLoop < numMatrices; matrixLoop++ )
589     {
590         double temp = GLCM->descriptors[ matrixLoop ][ descriptor ];
591         average += temp;
592         squareSum += temp*temp;
593     }
594 
595     average /= numMatrices;
596 
597     if( _average )
598         *_average = average;
599 
600     if( _standardDeviation )
601         *_standardDeviation = sqrt( (squareSum - average*average*numMatrices)/(numMatrices-1));
602 
603     __END__;
604 }
605 
606 
607 CV_IMPL IplImage*
cvCreateGLCMImage(CvGLCM * GLCM,int step)608 cvCreateGLCMImage( CvGLCM* GLCM, int step )
609 {
610     IplImage* dest = 0;
611 
612     CV_FUNCNAME( "cvCreateGLCMImage" );
613 
614     __BEGIN__;
615 
616     float* destData;
617     int sideLoop1, sideLoop2;
618 
619     if( !GLCM )
620         CV_ERROR( CV_StsNullPtr, "" );
621 
622     if( !(GLCM->matrices) )
623         CV_ERROR( CV_StsNullPtr, "Matrices are not allocated" );
624 
625     if( (unsigned)step >= (unsigned)(GLCM->numMatrices) )
626         CV_ERROR( CV_StsOutOfRange, "The step index is out of range" );
627 
628     dest = cvCreateImage( cvSize( GLCM->matrixSideLength, GLCM->matrixSideLength ), IPL_DEPTH_32F, 1 );
629     destData = (float*)(dest->imageData);
630 
631     for( sideLoop1 = 0; sideLoop1 < GLCM->matrixSideLength;
632                         sideLoop1++, (float*&)destData += dest->widthStep )
633     {
634         for( sideLoop2=0; sideLoop2 < GLCM->matrixSideLength; sideLoop2++ )
635         {
636             double matrixValue = GLCM->matrices[step][sideLoop1][sideLoop2];
637             destData[ sideLoop2 ] = (float)matrixValue;
638         }
639     }
640 
641     __END__;
642 
643     if( cvGetErrStatus() < 0 )
644         cvReleaseImage( &dest );
645 
646     return dest;
647 }
648 
649