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 #include "_cv.h"
42 
43 #define _CV_ACOS_TABLE_SIZE  513
44 
45 static const float icv_acos_table[_CV_ACOS_TABLE_SIZE] = {
46     3.14159265f, 3.05317551f, 3.01651113f, 2.98834964f, 2.96458497f, 2.94362719f,
47     2.92466119f, 2.90720289f, 2.89093699f, 2.87564455f, 2.86116621f, 2.84738169f,
48     2.83419760f, 2.82153967f, 2.80934770f, 2.79757211f, 2.78617145f, 2.77511069f,
49     2.76435988f, 2.75389319f, 2.74368816f, 2.73372510f, 2.72398665f, 2.71445741f,
50     2.70512362f, 2.69597298f, 2.68699438f, 2.67817778f, 2.66951407f, 2.66099493f,
51     2.65261279f, 2.64436066f, 2.63623214f, 2.62822133f, 2.62032277f, 2.61253138f,
52     2.60484248f, 2.59725167f, 2.58975488f, 2.58234828f, 2.57502832f, 2.56779164f,
53     2.56063509f, 2.55355572f, 2.54655073f, 2.53961750f, 2.53275354f, 2.52595650f,
54     2.51922417f, 2.51255441f, 2.50594525f, 2.49939476f, 2.49290115f, 2.48646269f,
55     2.48007773f, 2.47374472f, 2.46746215f, 2.46122860f, 2.45504269f, 2.44890314f,
56     2.44280867f, 2.43675809f, 2.43075025f, 2.42478404f, 2.41885841f, 2.41297232f,
57     2.40712480f, 2.40131491f, 2.39554173f, 2.38980439f, 2.38410204f, 2.37843388f,
58     2.37279910f, 2.36719697f, 2.36162673f, 2.35608768f, 2.35057914f, 2.34510044f,
59     2.33965094f, 2.33423003f, 2.32883709f, 2.32347155f, 2.31813284f, 2.31282041f,
60     2.30753373f, 2.30227228f, 2.29703556f, 2.29182309f, 2.28663439f, 2.28146900f,
61     2.27632647f, 2.27120637f, 2.26610827f, 2.26103177f, 2.25597646f, 2.25094195f,
62     2.24592786f, 2.24093382f, 2.23595946f, 2.23100444f, 2.22606842f, 2.22115104f,
63     2.21625199f, 2.21137096f, 2.20650761f, 2.20166166f, 2.19683280f, 2.19202074f,
64     2.18722520f, 2.18244590f, 2.17768257f, 2.17293493f, 2.16820274f, 2.16348574f,
65     2.15878367f, 2.15409630f, 2.14942338f, 2.14476468f, 2.14011997f, 2.13548903f,
66     2.13087163f, 2.12626757f, 2.12167662f, 2.11709859f, 2.11253326f, 2.10798044f,
67     2.10343994f, 2.09891156f, 2.09439510f, 2.08989040f, 2.08539725f, 2.08091550f,
68     2.07644495f, 2.07198545f, 2.06753681f, 2.06309887f, 2.05867147f, 2.05425445f,
69     2.04984765f, 2.04545092f, 2.04106409f, 2.03668703f, 2.03231957f, 2.02796159f,
70     2.02361292f, 2.01927344f, 2.01494300f, 2.01062146f, 2.00630870f, 2.00200457f,
71     1.99770895f, 1.99342171f, 1.98914271f, 1.98487185f, 1.98060898f, 1.97635399f,
72     1.97210676f, 1.96786718f, 1.96363511f, 1.95941046f, 1.95519310f, 1.95098292f,
73     1.94677982f, 1.94258368f, 1.93839439f, 1.93421185f, 1.93003595f, 1.92586659f,
74     1.92170367f, 1.91754708f, 1.91339673f, 1.90925250f, 1.90511432f, 1.90098208f,
75     1.89685568f, 1.89273503f, 1.88862003f, 1.88451060f, 1.88040664f, 1.87630806f,
76     1.87221477f, 1.86812668f, 1.86404371f, 1.85996577f, 1.85589277f, 1.85182462f,
77     1.84776125f, 1.84370256f, 1.83964848f, 1.83559892f, 1.83155381f, 1.82751305f,
78     1.82347658f, 1.81944431f, 1.81541617f, 1.81139207f, 1.80737194f, 1.80335570f,
79     1.79934328f, 1.79533460f, 1.79132959f, 1.78732817f, 1.78333027f, 1.77933581f,
80     1.77534473f, 1.77135695f, 1.76737240f, 1.76339101f, 1.75941271f, 1.75543743f,
81     1.75146510f, 1.74749565f, 1.74352900f, 1.73956511f, 1.73560389f, 1.73164527f,
82     1.72768920f, 1.72373560f, 1.71978441f, 1.71583556f, 1.71188899f, 1.70794462f,
83     1.70400241f, 1.70006228f, 1.69612416f, 1.69218799f, 1.68825372f, 1.68432127f,
84     1.68039058f, 1.67646160f, 1.67253424f, 1.66860847f, 1.66468420f, 1.66076139f,
85     1.65683996f, 1.65291986f, 1.64900102f, 1.64508338f, 1.64116689f, 1.63725148f,
86     1.63333709f, 1.62942366f, 1.62551112f, 1.62159943f, 1.61768851f, 1.61377831f,
87     1.60986877f, 1.60595982f, 1.60205142f, 1.59814349f, 1.59423597f, 1.59032882f,
88     1.58642196f, 1.58251535f, 1.57860891f, 1.57470259f, 1.57079633f, 1.56689007f,
89     1.56298375f, 1.55907731f, 1.55517069f, 1.55126383f, 1.54735668f, 1.54344917f,
90     1.53954124f, 1.53563283f, 1.53172389f, 1.52781434f, 1.52390414f, 1.51999323f,
91     1.51608153f, 1.51216900f, 1.50825556f, 1.50434117f, 1.50042576f, 1.49650927f,
92     1.49259163f, 1.48867280f, 1.48475270f, 1.48083127f, 1.47690845f, 1.47298419f,
93     1.46905841f, 1.46513106f, 1.46120207f, 1.45727138f, 1.45333893f, 1.44940466f,
94     1.44546850f, 1.44153038f, 1.43759024f, 1.43364803f, 1.42970367f, 1.42575709f,
95     1.42180825f, 1.41785705f, 1.41390346f, 1.40994738f, 1.40598877f, 1.40202755f,
96     1.39806365f, 1.39409701f, 1.39012756f, 1.38615522f, 1.38217994f, 1.37820164f,
97     1.37422025f, 1.37023570f, 1.36624792f, 1.36225684f, 1.35826239f, 1.35426449f,
98     1.35026307f, 1.34625805f, 1.34224937f, 1.33823695f, 1.33422072f, 1.33020059f,
99     1.32617649f, 1.32214834f, 1.31811607f, 1.31407960f, 1.31003885f, 1.30599373f,
100     1.30194417f, 1.29789009f, 1.29383141f, 1.28976803f, 1.28569989f, 1.28162688f,
101     1.27754894f, 1.27346597f, 1.26937788f, 1.26528459f, 1.26118602f, 1.25708205f,
102     1.25297262f, 1.24885763f, 1.24473698f, 1.24061058f, 1.23647833f, 1.23234015f,
103     1.22819593f, 1.22404557f, 1.21988898f, 1.21572606f, 1.21155670f, 1.20738080f,
104     1.20319826f, 1.19900898f, 1.19481283f, 1.19060973f, 1.18639955f, 1.18218219f,
105     1.17795754f, 1.17372548f, 1.16948589f, 1.16523866f, 1.16098368f, 1.15672081f,
106     1.15244994f, 1.14817095f, 1.14388370f, 1.13958808f, 1.13528396f, 1.13097119f,
107     1.12664966f, 1.12231921f, 1.11797973f, 1.11363107f, 1.10927308f, 1.10490563f,
108     1.10052856f, 1.09614174f, 1.09174500f, 1.08733820f, 1.08292118f, 1.07849378f,
109     1.07405585f, 1.06960721f, 1.06514770f, 1.06067715f, 1.05619540f, 1.05170226f,
110     1.04719755f, 1.04268110f, 1.03815271f, 1.03361221f, 1.02905939f, 1.02449407f,
111     1.01991603f, 1.01532509f, 1.01072102f, 1.00610363f, 1.00147268f, 0.99682798f,
112     0.99216928f, 0.98749636f, 0.98280898f, 0.97810691f, 0.97338991f, 0.96865772f,
113     0.96391009f, 0.95914675f, 0.95436745f, 0.94957191f, 0.94475985f, 0.93993099f,
114     0.93508504f, 0.93022170f, 0.92534066f, 0.92044161f, 0.91552424f, 0.91058821f,
115     0.90563319f, 0.90065884f, 0.89566479f, 0.89065070f, 0.88561619f, 0.88056088f,
116     0.87548438f, 0.87038629f, 0.86526619f, 0.86012366f, 0.85495827f, 0.84976956f,
117     0.84455709f, 0.83932037f, 0.83405893f, 0.82877225f, 0.82345981f, 0.81812110f,
118     0.81275556f, 0.80736262f, 0.80194171f, 0.79649221f, 0.79101352f, 0.78550497f,
119     0.77996593f, 0.77439569f, 0.76879355f, 0.76315878f, 0.75749061f, 0.75178826f,
120     0.74605092f, 0.74027775f, 0.73446785f, 0.72862033f, 0.72273425f, 0.71680861f,
121     0.71084240f, 0.70483456f, 0.69878398f, 0.69268952f, 0.68654996f, 0.68036406f,
122     0.67413051f, 0.66784794f, 0.66151492f, 0.65512997f, 0.64869151f, 0.64219789f,
123     0.63564741f, 0.62903824f, 0.62236849f, 0.61563615f, 0.60883911f, 0.60197515f,
124     0.59504192f, 0.58803694f, 0.58095756f, 0.57380101f, 0.56656433f, 0.55924437f,
125     0.55183778f, 0.54434099f, 0.53675018f, 0.52906127f, 0.52126988f, 0.51337132f,
126     0.50536051f, 0.49723200f, 0.48897987f, 0.48059772f, 0.47207859f, 0.46341487f,
127     0.45459827f, 0.44561967f, 0.43646903f, 0.42713525f, 0.41760600f, 0.40786755f,
128     0.39790449f, 0.38769946f, 0.37723277f, 0.36648196f, 0.35542120f, 0.34402054f,
129     0.33224495f, 0.32005298f, 0.30739505f, 0.29421096f, 0.28042645f, 0.26594810f,
130     0.25065566f, 0.23438976f, 0.21693146f, 0.19796546f, 0.17700769f, 0.15324301f,
131     0.12508152f, 0.08841715f, 0.00000000f
132 };
133 
134 
135 /*F///////////////////////////////////////////////////////////////////////////////////////
136 //    Name: icvCalcPGH
137 //    Purpose:
138 //      Calculates PGH(pairwise geometric histogram) for contour given.
139 //    Context:
140 //    Parameters:
141 //      contour  - pointer to input contour object.
142 //      pgh      - output histogram
143 //      ang_dim  - number of angle bins (vertical size of histogram)
144 //      dist_dim - number of distance bins (horizontal size of histogram)
145 //    Returns:
146 //      CV_OK or error code
147 //    Notes:
148 //F*/
149 static CvStatus
icvCalcPGH(const CvSeq * contour,float * pgh,int angle_dim,int dist_dim)150 icvCalcPGH( const CvSeq * contour, float *pgh, int angle_dim, int dist_dim )
151 {
152     char local_buffer[(1 << 14) + 32];
153     float *local_buffer_ptr = (float *)cvAlignPtr(local_buffer,32);
154     float *buffer = local_buffer_ptr;
155     double angle_scale = (angle_dim - 0.51) / icv_acos_table[0];
156     double dist_scale = DBL_EPSILON;
157     int buffer_size;
158     int i, count, pass;
159     int *pghi = (int *) pgh;
160     int hist_size = angle_dim * dist_dim;
161     CvSeqReader reader1, reader2;       /* external and internal readers */
162 
163     if( !contour || !pgh )
164         return CV_NULLPTR_ERR;
165 
166     if( angle_dim <= 0 || angle_dim > 180 || dist_dim <= 0 )
167         return CV_BADRANGE_ERR;
168 
169     if( !CV_IS_SEQ_POLYGON( contour ))
170         return CV_BADFLAG_ERR;
171 
172     memset( pgh, 0, hist_size * sizeof( pgh[0] ));
173 
174     count = contour->total;
175 
176     /* allocate buffer for distances */
177     buffer_size = count * sizeof( float );
178 
179     if( buffer_size > (int)sizeof(local_buffer) - 32 )
180     {
181         buffer = (float *) cvAlloc( buffer_size );
182         if( !buffer )
183             return CV_OUTOFMEM_ERR;
184     }
185 
186     cvStartReadSeq( contour, &reader1, 0 );
187     cvStartReadSeq( contour, &reader2, 0 );
188 
189     /* calc & store squared edge lengths, calculate maximal distance between edges */
190     for( i = 0; i < count; i++ )
191     {
192         CvPoint pt1, pt2;
193         double dx, dy;
194 
195         CV_READ_EDGE( pt1, pt2, reader1 );
196 
197         dx = pt2.x - pt1.x;
198         dy = pt2.y - pt1.y;
199         buffer[i] = (float)(1./sqrt(dx * dx + dy * dy));
200     }
201 
202     /*
203        do 2 passes.
204        First calculates maximal distance.
205        Second calculates histogram itself.
206      */
207     for( pass = 1; pass <= 2; pass++ )
208     {
209         double dist_coeff = 0, angle_coeff = 0;
210 
211         /* run external loop */
212         for( i = 0; i < count; i++ )
213         {
214             CvPoint pt1, pt2;
215             int dx, dy;
216             int dist = 0;
217 
218             CV_READ_EDGE( pt1, pt2, reader1 );
219 
220             dx = pt2.x - pt1.x;
221             dy = pt2.y - pt1.y;
222 
223             if( (dx | dy) != 0 )
224             {
225                 int j;
226 
227                 if( pass == 2 )
228                 {
229                     dist_coeff = buffer[i] * dist_scale;
230                     angle_coeff = buffer[i] * (_CV_ACOS_TABLE_SIZE / 2);
231                 }
232 
233                 /* run internal loop (for current edge) */
234                 for( j = 0; j < count; j++ )
235                 {
236                     CvPoint pt3, pt4;
237 
238                     CV_READ_EDGE( pt3, pt4, reader2 );
239 
240                     if( i != j )        /* process edge pair */
241                     {
242                         int d1 = (pt3.y - pt1.y) * dx - (pt3.x - pt1.x) * dy;
243                         int d2 = (pt4.y - pt1.y) * dx - (pt2.x - pt1.x) * dy;
244                         int cross_flag;
245                         int *hist_row = 0;
246 
247                         if( pass == 2 )
248                         {
249                             int dp = (pt4.x - pt3.x) * dx + (pt4.y - pt3.y) * dy;
250 
251                             dp = cvRound( dp * angle_coeff * buffer[j] ) +
252                                 (_CV_ACOS_TABLE_SIZE / 2);
253                             dp = MAX( dp, 0 );
254                             dp = MIN( dp, _CV_ACOS_TABLE_SIZE - 1 );
255                             hist_row = pghi + dist_dim *
256                                 cvRound( icv_acos_table[dp] * angle_scale );
257 
258                             d1 = cvRound( d1 * dist_coeff );
259                             d2 = cvRound( d2 * dist_coeff );
260                         }
261 
262                         cross_flag = (d1 ^ d2) < 0;
263 
264                         d1 = CV_IABS( d1 );
265                         d2 = CV_IABS( d2 );
266 
267                         if( pass == 2 )
268                         {
269                             if( d1 >= dist_dim )
270                                 d1 = dist_dim - 1;
271                             if( d2 >= dist_dim )
272                                 d2 = dist_dim - 1;
273 
274                             if( !cross_flag )
275                             {
276                                 if( d1 > d2 )   /* make d1 <= d2 */
277                                 {
278                                     d1 ^= d2;
279                                     d2 ^= d1;
280                                     d1 ^= d2;
281                                 }
282 
283                                 for( ; d1 <= d2; d1++ )
284                                     hist_row[d1]++;
285                             }
286                             else
287                             {
288                                 for( ; d1 >= 0; d1-- )
289                                     hist_row[d1]++;
290                                 for( ; d2 >= 0; d2-- )
291                                     hist_row[d2]++;
292                             }
293                         }
294                         else    /* 1st pass */
295                         {
296                             d1 = CV_IMAX( d1, d2 );
297                             dist = CV_IMAX( dist, d1 );
298                         }
299                     }           /* end of processing of edge pair */
300 
301                 }               /* end of internal loop */
302 
303                 if( pass == 1 )
304                 {
305                     double scale = dist * buffer[i];
306 
307                     dist_scale = MAX( dist_scale, scale );
308                 }
309             }
310         }                       /* end of external loop */
311 
312         if( pass == 1 )
313         {
314             dist_scale = (dist_dim - 0.51) / dist_scale;
315         }
316 
317     }                           /* end of pass on loops */
318 
319 
320     /* convert hist to floats */
321     for( i = 0; i < hist_size; i++ )
322     {
323         ((float *) pghi)[i] = (float) pghi[i];
324     }
325 
326     if( buffer != local_buffer_ptr )
327         cvFree( &buffer );
328 
329     return CV_OK;
330 }
331 
332 
333 CV_IMPL void
cvCalcPGH(const CvSeq * contour,CvHistogram * hist)334 cvCalcPGH( const CvSeq * contour, CvHistogram * hist )
335 {
336     CV_FUNCNAME( "cvCalcPGH" );
337 
338     __BEGIN__;
339 
340     int size[CV_MAX_DIM];
341     int dims;
342 
343     if( !CV_IS_HIST(hist))
344         CV_ERROR( CV_StsBadArg, "The histogram header is invalid " );
345 
346     if( CV_IS_SPARSE_HIST( hist ))
347         CV_ERROR( CV_StsUnsupportedFormat, "Sparse histogram are not supported" );
348 
349     dims = cvGetDims( hist->bins, size );
350 
351     if( dims != 2 )
352         CV_ERROR( CV_StsBadSize, "The histogram must be two-dimensional" );
353 
354     if( !CV_IS_SEQ_POLYGON( contour ) || CV_SEQ_ELTYPE( contour ) != CV_32SC2 )
355         CV_ERROR( CV_StsUnsupportedFormat, "The contour is not valid or the point type is not supported" );
356 
357     IPPI_CALL( icvCalcPGH( contour, ((CvMatND*)(hist->bins))->data.fl, size[0], size[1] ));
358 
359     __END__;
360 }
361 
362 
363 /* End of file. */
364