/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// Intel License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000, Intel Corporation, all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of Intel Corporation may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "_cxcore.h"
/****************************************************************************************\
* Find sum of pixels in the ROI *
\****************************************************************************************/
#define ICV_SUM_COI_CASE( __op__, len, cn ) \
for( ; x <= (len) - 4*(cn); x += 4*(cn) ) \
s0 += __op__(src[x]) + __op__(src[x+(cn)]) + \
__op__(src[x+(cn)*2]) + __op__(src[x+(cn)*3]);\
\
for( ; x < (len); x += (cn) ) \
s0 += __op__(src[x]);
#define ICV_SUM_CASE_C1( __op__, len ) \
ICV_SUM_COI_CASE( __op__, len, 1 )
#define ICV_SUM_CASE_C2( __op__, len ) \
for( ; x <= (len) - 8; x += 8 ) \
{ \
s0 += __op__(src[x]) + __op__(src[x+2]) + \
__op__(src[x+4]) + __op__(src[x+6]); \
s1 += __op__(src[x+1]) + __op__(src[x+3]) + \
__op__(src[x+5]) + __op__(src[x+7]); \
} \
\
for( ; x < (len); x += 2 ) \
{ \
s0 += __op__(src[x]); \
s1 += __op__(src[x+1]); \
}
#define ICV_SUM_CASE_C3( __op__, len ) \
for( ; x <= (len) - 12; x += 12 ) \
{ \
s0 += __op__(src[x]) + __op__(src[x+3]) + \
__op__(src[x+6]) + __op__(src[x+9]); \
s1 += __op__(src[x+1]) + __op__(src[x+4]) + \
__op__(src[x+7]) + __op__(src[x+10]); \
s2 += __op__(src[x+2]) + __op__(src[x+5]) + \
__op__(src[x+8]) + __op__(src[x+11]); \
} \
\
for( ; x < (len); x += 3 ) \
{ \
s0 += __op__(src[x]); \
s1 += __op__(src[x+1]); \
s2 += __op__(src[x+2]); \
}
#define ICV_SUM_CASE_C4( __op__, len ) \
for( ; x <= (len) - 16; x += 16 ) \
{ \
s0 += __op__(src[x]) + __op__(src[x+4]) + \
__op__(src[x+8]) + __op__(src[x+12]); \
s1 += __op__(src[x+1]) + __op__(src[x+5]) + \
__op__(src[x+9]) + __op__(src[x+13]); \
s2 += __op__(src[x+2]) + __op__(src[x+6]) + \
__op__(src[x+10]) + __op__(src[x+14]); \
s3 += __op__(src[x+3]) + __op__(src[x+7]) + \
__op__(src[x+11]) + __op__(src[x+15]); \
} \
\
for( ; x < (len); x += 4 ) \
{ \
s0 += __op__(src[x]); \
s1 += __op__(src[x+1]); \
s2 += __op__(src[x+2]); \
s3 += __op__(src[x+3]); \
}
////////////////////////////////////// entry macros //////////////////////////////////////
#define ICV_SUM_ENTRY_COMMON() \
step /= sizeof(src[0])
#define ICV_SUM_ENTRY_C1( sumtype ) \
sumtype s0 = 0; \
ICV_SUM_ENTRY_COMMON()
#define ICV_SUM_ENTRY_C2( sumtype ) \
sumtype s0 = 0, s1 = 0; \
ICV_SUM_ENTRY_COMMON()
#define ICV_SUM_ENTRY_C3( sumtype ) \
sumtype s0 = 0, s1 = 0, s2 = 0; \
ICV_SUM_ENTRY_COMMON()
#define ICV_SUM_ENTRY_C4( sumtype ) \
sumtype s0 = 0, s1 = 0, s2 = 0, s3 = 0; \
ICV_SUM_ENTRY_COMMON()
#define ICV_SUM_ENTRY_BLOCK_COMMON( block_size ) \
int remaining = block_size; \
ICV_SUM_ENTRY_COMMON()
#define ICV_SUM_ENTRY_BLOCK_C1( sumtype, worktype, block_size ) \
sumtype sum0 = 0; \
worktype s0 = 0; \
ICV_SUM_ENTRY_BLOCK_COMMON( block_size )
#define ICV_SUM_ENTRY_BLOCK_C2( sumtype, worktype, block_size ) \
sumtype sum0 = 0, sum1 = 0; \
worktype s0 = 0, s1 = 0; \
ICV_SUM_ENTRY_BLOCK_COMMON( block_size )
#define ICV_SUM_ENTRY_BLOCK_C3( sumtype, worktype, block_size ) \
sumtype sum0 = 0, sum1 = 0, sum2 = 0; \
worktype s0 = 0, s1 = 0, s2 = 0; \
ICV_SUM_ENTRY_BLOCK_COMMON( block_size )
#define ICV_SUM_ENTRY_BLOCK_C4( sumtype, worktype, block_size ) \
sumtype sum0 = 0, sum1 = 0, sum2 = 0, sum3 = 0; \
worktype s0 = 0, s1 = 0, s2 = 0, s3 = 0; \
ICV_SUM_ENTRY_BLOCK_COMMON( block_size )
/////////////////////////////////////// exit macros //////////////////////////////////////
#define ICV_SUM_EXIT_C1( tmp, sumtype ) \
sum[0] = (sumtype)tmp##0
#define ICV_SUM_EXIT_C2( tmp, sumtype ) \
sum[0] = (sumtype)tmp##0; \
sum[1] = (sumtype)tmp##1;
#define ICV_SUM_EXIT_C3( tmp, sumtype ) \
sum[0] = (sumtype)tmp##0; \
sum[1] = (sumtype)tmp##1; \
sum[2] = (sumtype)tmp##2;
#define ICV_SUM_EXIT_C4( tmp, sumtype ) \
sum[0] = (sumtype)tmp##0; \
sum[1] = (sumtype)tmp##1; \
sum[2] = (sumtype)tmp##2; \
sum[3] = (sumtype)tmp##3;
#define ICV_SUM_EXIT_BLOCK_C1( sumtype ) \
sum0 += s0; \
ICV_SUM_EXIT_C1( sum, sumtype )
#define ICV_SUM_EXIT_BLOCK_C2( sumtype ) \
sum0 += s0; sum1 += s1; \
ICV_SUM_EXIT_C2( sum, sumtype )
#define ICV_SUM_EXIT_BLOCK_C3( sumtype ) \
sum0 += s0; sum1 += s1; \
sum2 += s2; \
ICV_SUM_EXIT_C3( sum, sumtype )
#define ICV_SUM_EXIT_BLOCK_C4( sumtype ) \
sum0 += s0; sum1 += s1; \
sum2 += s2; sum3 += s3; \
ICV_SUM_EXIT_C4( sum, sumtype )
////////////////////////////////////// update macros /////////////////////////////////////
#define ICV_SUM_UPDATE_COMMON( block_size ) \
remaining = block_size
#define ICV_SUM_UPDATE_C1( block_size ) \
ICV_SUM_UPDATE_COMMON( block_size ); \
sum0 += s0; \
s0 = 0
#define ICV_SUM_UPDATE_C2( block_size ) \
ICV_SUM_UPDATE_COMMON( block_size ); \
sum0 += s0; sum1 += s1; \
s0 = s1 = 0
#define ICV_SUM_UPDATE_C3( block_size ) \
ICV_SUM_UPDATE_COMMON( block_size ); \
sum0 += s0; sum1 += s1; sum2 += s2; \
s0 = s1 = s2 = 0
#define ICV_SUM_UPDATE_C4( block_size ) \
ICV_SUM_UPDATE_COMMON( block_size ); \
sum0 += s0; sum1 += s1; \
sum2 += s2; sum3 += s3; \
s0 = s1 = s2 = s3 = 0
#define ICV_DEF_SUM_NOHINT_BLOCK_FUNC_2D( name, flavor, cn, \
__op__, arrtype, sumtype_final, sumtype, worktype, block_size )\
IPCVAPI_IMPL(CvStatus, icv##name##_##flavor##_C##cn##R,( \
const arrtype* src, int step, CvSize size, \
sumtype_final* sum ), (src, step, size, sum) ) \
{ \
ICV_SUM_ENTRY_BLOCK_C##cn(sumtype,worktype,(block_size)*(cn)); \
size.width *= cn; \
\
for( ; size.height--; src += step ) \
{ \
int x = 0; \
while( x < size.width ) \
{ \
int limit = MIN( remaining, size.width - x ); \
remaining -= limit; \
limit += x; \
ICV_SUM_CASE_C##cn( __op__, limit ); \
if( remaining == 0 ) \
{ \
ICV_SUM_UPDATE_C##cn( (block_size)*(cn) ); \
} \
} \
} \
\
ICV_SUM_EXIT_BLOCK_C##cn( sumtype_final ); \
return CV_OK; \
}
#define ICV_DEF_SUM_NOHINT_FUNC_2D( name, flavor, cn, \
__op__, arrtype, sumtype_final, sumtype, worktype, block_size )\
IPCVAPI_IMPL(CvStatus, icv##name##_##flavor##_C##cn##R,( \
const arrtype* src, int step, CvSize size, \
sumtype_final* sum ), (src, step, size, sum) ) \
{ \
ICV_SUM_ENTRY_C##cn( sumtype ); \
size.width *= cn; \
\
for( ; size.height--; src += step ) \
{ \
int x = 0; \
ICV_SUM_CASE_C##cn( __op__, size.width ); \
} \
\
ICV_SUM_EXIT_C##cn( s, sumtype_final ); \
return CV_OK; \
}
#define ICV_DEF_SUM_HINT_FUNC_2D( name, flavor, cn, \
__op__, arrtype, sumtype_final, sumtype, worktype, block_size )\
IPCVAPI_IMPL(CvStatus, icv##name##_##flavor##_C##cn##R,( \
const arrtype* src, int step, CvSize size, \
sumtype_final* sum, CvHintAlgorithm /*hint*/ ), \
(src, step, size, sum, cvAlgHintAccurate) ) \
{ \
ICV_SUM_ENTRY_C##cn( sumtype ); \
size.width *= cn; \
\
for( ; size.height--; src += step ) \
{ \
int x = 0; \
ICV_SUM_CASE_C##cn( __op__, size.width ); \
} \
\
ICV_SUM_EXIT_C##cn( s, sumtype_final ); \
return CV_OK; \
}
#define ICV_DEF_SUM_NOHINT_BLOCK_FUNC_2D_COI( name, flavor, \
__op__, arrtype, sumtype_final, sumtype, worktype, block_size )\
static CvStatus CV_STDCALL icv##name##_##flavor##_CnCR( \
const arrtype* src, int step, CvSize size, int cn, \
int coi, sumtype_final* sum ) \
{ \
ICV_SUM_ENTRY_BLOCK_C1(sumtype,worktype,(block_size)*(cn)); \
size.width *= cn; \
src += coi - 1; \
\
for( ; size.height--; src += step ) \
{ \
int x = 0; \
while( x < size.width ) \
{ \
int limit = MIN( remaining, size.width - x ); \
remaining -= limit; \
limit += x; \
ICV_SUM_COI_CASE( __op__, limit, cn ); \
if( remaining == 0 ) \
{ \
ICV_SUM_UPDATE_C1( (block_size)*(cn) ); \
} \
} \
} \
\
ICV_SUM_EXIT_BLOCK_C1( sumtype_final ); \
return CV_OK; \
}
#define ICV_DEF_SUM_NOHINT_FUNC_2D_COI( name, flavor, \
__op__, arrtype, sumtype_final, sumtype, worktype, block_size )\
static CvStatus CV_STDCALL icv##name##_##flavor##_CnCR( \
const arrtype* src, int step, CvSize size, int cn, \
int coi, sumtype_final* sum ) \
{ \
ICV_SUM_ENTRY_C1( sumtype ); \
size.width *= cn; \
src += coi - 1; \
\
for( ; size.height--; src += step ) \
{ \
int x = 0; \
ICV_SUM_COI_CASE( __op__, size.width, cn ); \
} \
\
ICV_SUM_EXIT_C1( s, sumtype_final ); \
return CV_OK; \
}
#define ICV_DEF_SUM_ALL( name, flavor, __op__, arrtype, sumtype_final, sumtype, \
worktype, hintp_type, nohint_type, block_size ) \
ICV_DEF_SUM_##hintp_type##_FUNC_2D( name, flavor, 1, __op__, arrtype, \
sumtype_final, sumtype, worktype, block_size ) \
ICV_DEF_SUM_##hintp_type##_FUNC_2D( name, flavor, 2, __op__, arrtype, \
sumtype_final, sumtype, worktype, block_size ) \
ICV_DEF_SUM_##hintp_type##_FUNC_2D( name, flavor, 3, __op__, arrtype, \
sumtype_final, sumtype, worktype, block_size ) \
ICV_DEF_SUM_##hintp_type##_FUNC_2D( name, flavor, 4, __op__, arrtype, \
sumtype_final, sumtype, worktype, block_size ) \
ICV_DEF_SUM_##nohint_type##_FUNC_2D_COI( name, flavor, __op__, arrtype, \
sumtype_final, sumtype, worktype, block_size )
ICV_DEF_SUM_ALL( Sum, 8u, CV_NOP, uchar, double, int64, unsigned,
NOHINT_BLOCK, NOHINT_BLOCK, 1 << 24 )
ICV_DEF_SUM_ALL( Sum, 16u, CV_NOP, ushort, double, int64, unsigned,
NOHINT_BLOCK, NOHINT_BLOCK, 1 << 16 )
ICV_DEF_SUM_ALL( Sum, 16s, CV_NOP, short, double, int64, int,
NOHINT_BLOCK, NOHINT_BLOCK, 1 << 16 )
ICV_DEF_SUM_ALL( Sum, 32s, CV_NOP, int, double, double, double, NOHINT, NOHINT, 0 )
ICV_DEF_SUM_ALL( Sum, 32f, CV_NOP, float, double, double, double, HINT, NOHINT, 0 )
ICV_DEF_SUM_ALL( Sum, 64f, CV_NOP, double, double, double, double, NOHINT, NOHINT, 0 )
#define icvSum_8s_C1R 0
#define icvSum_8s_C2R 0
#define icvSum_8s_C3R 0
#define icvSum_8s_C4R 0
#define icvSum_8s_CnCR 0
CV_DEF_INIT_BIG_FUNC_TAB_2D( Sum, R )
CV_DEF_INIT_FUNC_TAB_2D( Sum, CnCR )
CV_IMPL CvScalar
cvSum( const CvArr* arr )
{
static CvBigFuncTable sum_tab;
static CvFuncTable sumcoi_tab;
static int inittab = 0;
CvScalar sum = {{0,0,0,0}};
CV_FUNCNAME("cvSum");
__BEGIN__;
int type, coi = 0;
int mat_step;
CvSize size;
CvMat stub, *mat = (CvMat*)arr;
if( !inittab )
{
icvInitSumRTable( &sum_tab );
icvInitSumCnCRTable( &sumcoi_tab );
inittab = 1;
}
if( !CV_IS_MAT(mat) )
{
if( CV_IS_MATND(mat) )
{
void* matnd = (void*)mat;
CvMatND nstub;
CvNArrayIterator iterator;
int pass_hint;
CV_CALL( cvInitNArrayIterator( 1, &matnd, 0, &nstub, &iterator ));
type = CV_MAT_TYPE(iterator.hdr[0]->type);
if( CV_MAT_CN(type) > 4 )
CV_ERROR( CV_StsOutOfRange, "The input array must have at most 4 channels" );
pass_hint = CV_MAT_DEPTH(type) == CV_32F;
if( !pass_hint )
{
CvFunc2D_1A1P func = (CvFunc2D_1A1P)(sum_tab.fn_2d[type]);
if( !func )
CV_ERROR( CV_StsUnsupportedFormat, "" );
do
{
CvScalar temp = {{0,0,0,0}};
IPPI_CALL( func( iterator.ptr[0], CV_STUB_STEP,
iterator.size, temp.val ));
sum.val[0] += temp.val[0];
sum.val[1] += temp.val[1];
sum.val[2] += temp.val[2];
sum.val[3] += temp.val[3];
}
while( cvNextNArraySlice( &iterator ));
}
else
{
CvFunc2D_1A1P1I func = (CvFunc2D_1A1P1I)(sum_tab.fn_2d[type]);
if( !func )
CV_ERROR( CV_StsUnsupportedFormat, "" );
do
{
CvScalar temp = {{0,0,0,0}};
IPPI_CALL( func( iterator.ptr[0], CV_STUB_STEP,
iterator.size, temp.val, cvAlgHintAccurate ));
sum.val[0] += temp.val[0];
sum.val[1] += temp.val[1];
sum.val[2] += temp.val[2];
sum.val[3] += temp.val[3];
}
while( cvNextNArraySlice( &iterator ));
}
EXIT;
}
else
CV_CALL( mat = cvGetMat( mat, &stub, &coi ));
}
type = CV_MAT_TYPE(mat->type);
size = cvGetMatSize( mat );
mat_step = mat->step;
if( CV_IS_MAT_CONT( mat->type ))
{
size.width *= size.height;
if( size.width <= CV_MAX_INLINE_MAT_OP_SIZE )
{
if( type == CV_32FC1 )
{
float* data = mat->data.fl;
do
{
sum.val[0] += data[size.width - 1];
}
while( --size.width );
EXIT;
}
if( type == CV_64FC1 )
{
double* data = mat->data.db;
do
{
sum.val[0] += data[size.width - 1];
}
while( --size.width );
EXIT;
}
}
size.height = 1;
mat_step = CV_STUB_STEP;
}
if( CV_MAT_CN(type) == 1 || coi == 0 )
{
int pass_hint = CV_MAT_DEPTH(type) == CV_32F;
if( CV_MAT_CN(type) > 4 )
CV_ERROR( CV_StsOutOfRange, "The input array must have at most 4 channels" );
if( !pass_hint )
{
CvFunc2D_1A1P func = (CvFunc2D_1A1P)(sum_tab.fn_2d[type]);
if( !func )
CV_ERROR( CV_StsBadArg, cvUnsupportedFormat );
IPPI_CALL( func( mat->data.ptr, mat_step, size, sum.val ));
}
else
{
CvFunc2D_1A1P1I func = (CvFunc2D_1A1P1I)(sum_tab.fn_2d[type]);
if( !func )
CV_ERROR( CV_StsBadArg, cvUnsupportedFormat );
IPPI_CALL( func( mat->data.ptr, mat_step, size, sum.val, cvAlgHintAccurate ));
}
}
else
{
CvFunc2DnC_1A1P func = (CvFunc2DnC_1A1P)(sumcoi_tab.fn_2d[CV_MAT_DEPTH(type)]);
if( !func )
CV_ERROR( CV_StsBadArg, cvUnsupportedFormat );
IPPI_CALL( func( mat->data.ptr, mat_step, size,
CV_MAT_CN(type), coi, sum.val ));
}
__END__;
return sum;
}
#define ICV_DEF_NONZERO_ALL( flavor, __op__, arrtype ) \
ICV_DEF_SUM_NOHINT_FUNC_2D( CountNonZero, flavor, 1, __op__, \
arrtype, int, int, int, 0 ) \
ICV_DEF_SUM_NOHINT_FUNC_2D_COI( CountNonZero, flavor, __op__, \
arrtype, int, int, int, 0 )
#undef CV_NONZERO_DBL
#define CV_NONZERO_DBL(x) (((x) & CV_BIG_INT(0x7fffffffffffffff)) != 0)
ICV_DEF_NONZERO_ALL( 8u, CV_NONZERO, uchar )
ICV_DEF_NONZERO_ALL( 16s, CV_NONZERO, ushort )
ICV_DEF_NONZERO_ALL( 32s, CV_NONZERO, int )
ICV_DEF_NONZERO_ALL( 32f, CV_NONZERO_FLT, int )
ICV_DEF_NONZERO_ALL( 64f, CV_NONZERO_DBL, int64 )
#define icvCountNonZero_8s_C1R icvCountNonZero_8u_C1R
#define icvCountNonZero_8s_CnCR icvCountNonZero_8u_CnCR
#define icvCountNonZero_16u_C1R icvCountNonZero_16s_C1R
#define icvCountNonZero_16u_CnCR icvCountNonZero_16s_CnCR
CV_DEF_INIT_FUNC_TAB_2D( CountNonZero, C1R )
CV_DEF_INIT_FUNC_TAB_2D( CountNonZero, CnCR )
CV_IMPL int
cvCountNonZero( const CvArr* arr )
{
static CvFuncTable nz_tab;
static CvFuncTable nzcoi_tab;
static int inittab = 0;
int count = 0;
CV_FUNCNAME("cvCountNonZero");
__BEGIN__;
int type, coi = 0;
int mat_step;
CvSize size;
CvMat stub, *mat = (CvMat*)arr;
if( !inittab )
{
icvInitCountNonZeroC1RTable( &nz_tab );
icvInitCountNonZeroCnCRTable( &nzcoi_tab );
inittab = 1;
}
if( !CV_IS_MAT(mat) )
{
if( CV_IS_MATND(mat) )
{
void* matnd = (void*)arr;
CvMatND nstub;
CvNArrayIterator iterator;
CvFunc2D_1A1P func;
CV_CALL( cvInitNArrayIterator( 1, &matnd, 0, &nstub, &iterator ));
type = CV_MAT_TYPE(iterator.hdr[0]->type);
if( CV_MAT_CN(type) != 1 )
CV_ERROR( CV_BadNumChannels,
"Only single-channel array are supported here" );
func = (CvFunc2D_1A1P)(nz_tab.fn_2d[CV_MAT_DEPTH(type)]);
if( !func )
CV_ERROR( CV_StsUnsupportedFormat, "" );
do
{
int temp;
IPPI_CALL( func( iterator.ptr[0], CV_STUB_STEP,
iterator.size, &temp ));
count += temp;
}
while( cvNextNArraySlice( &iterator ));
EXIT;
}
else
CV_CALL( mat = cvGetMat( mat, &stub, &coi ));
}
type = CV_MAT_TYPE(mat->type);
size = cvGetMatSize( mat );
mat_step = mat->step;
if( CV_IS_MAT_CONT( mat->type ))
{
size.width *= size.height;
size.height = 1;
mat_step = CV_STUB_STEP;
}
if( CV_MAT_CN(type) == 1 || coi == 0 )
{
CvFunc2D_1A1P func = (CvFunc2D_1A1P)(nz_tab.fn_2d[CV_MAT_DEPTH(type)]);
if( CV_MAT_CN(type) != 1 )
CV_ERROR( CV_BadNumChannels,
"The function can handle only a single channel at a time (use COI)");
if( !func )
CV_ERROR( CV_StsBadArg, cvUnsupportedFormat );
IPPI_CALL( func( mat->data.ptr, mat_step, size, &count ));
}
else
{
CvFunc2DnC_1A1P func = (CvFunc2DnC_1A1P)(nzcoi_tab.fn_2d[CV_MAT_DEPTH(type)]);
if( !func )
CV_ERROR( CV_StsBadArg, cvUnsupportedFormat );
IPPI_CALL( func( mat->data.ptr, mat_step, size, CV_MAT_CN(type), coi, &count ));
}
__END__;
return count;
}
/****************************************************************************************\
* Reduce Matrix to Vector *
\****************************************************************************************/
#define ICV_ACC_ROWS_FUNC( name, flavor, arrtype, acctype, \
__op__, load_macro ) \
static CvStatus CV_STDCALL \
icv##name##Rows_##flavor##_C1R( const arrtype* src, int srcstep,\
acctype* dst, CvSize size ) \
{ \
int i, width = size.width; \
srcstep /= sizeof(src[0]); \
\
for( i = 0; i < width; i++ ) \
dst[i] = load_macro(src[i]); \
\
for( ; --size.height; ) \
{ \
src += srcstep; \
for( i = 0; i <= width - 4; i += 4 ) \
{ \
acctype s0 = load_macro(src[i]); \
acctype s1 = load_macro(src[i+1]); \
acctype a0 = dst[i], a1 = dst[i+1]; \
a0 = (acctype)__op__(a0,s0); a1 = (acctype)__op__(a1,s1); \
dst[i] = a0; dst[i+1] = a1; \
\
s0 = load_macro(src[i+2]); \
s1 = load_macro(src[i+3]); \
a0 = dst[i+2]; a1 = dst[i+3]; \
a0 = (acctype)__op__(a0,s0); a1 = (acctype)__op__(a1,s1); \
dst[i+2] = a0; dst[i+3] = a1; \
} \
\
for( ; i < width; i++ ) \
{ \
acctype s0 = load_macro(src[i]), a0 = dst[i]; \
a0 = (acctype)__op__(a0,s0); \
dst[i] = a0; \
} \
} \
\
return CV_OK; \
}
#define ICV_ACC_COLS_FUNC_C1( name, flavor, arrtype, worktype, acctype, __op__ )\
static CvStatus CV_STDCALL \
icv##name##Cols_##flavor##_C1R( const arrtype* src, int srcstep, \
acctype* dst, int dststep, CvSize size )\
{ \
int i, width = size.width; \
srcstep /= sizeof(src[0]); \
dststep /= sizeof(dst[0]); \
\
for( ; size.height--; src += srcstep, dst += dststep ) \
{ \
if( width == 1 ) \
dst[0] = (acctype)src[0]; \
else \
{ \
worktype a0 = src[0], a1 = src[1]; \
for( i = 2; i <= width - 4; i += 4 ) \
{ \
worktype s0 = src[i], s1 = src[i+1]; \
a0 = __op__(a0, s0); \
a1 = __op__(a1, s1); \
s0 = src[i+2]; s1 = src[i+3]; \
a0 = __op__(a0, s0); \
a1 = __op__(a1, s1); \
} \
\
for( ; i < width; i++ ) \
{ \
worktype s0 = src[i]; \
a0 = __op__(a0, s0); \
} \
a0 = __op__(a0, a1); \
dst[0] = (acctype)a0; \
} \
} \
\
return CV_OK; \
}
#define ICV_ACC_COLS_FUNC_C3( name, flavor, arrtype, worktype, acctype, __op__ ) \
static CvStatus CV_STDCALL \
icv##name##Cols_##flavor##_C3R( const arrtype* src, int srcstep, \
acctype* dst, int dststep, CvSize size )\
{ \
int i, width = size.width*3; \
srcstep /= sizeof(src[0]); \
dststep /= sizeof(dst[0]); \
\
for( ; size.height--; src += srcstep, dst += dststep ) \
{ \
worktype a0 = src[0], a1 = src[1], a2 = src[2]; \
for( i = 3; i < width; i += 3 ) \
{ \
worktype s0 = src[i], s1 = src[i+1], s2 = src[i+2]; \
a0 = __op__(a0, s0); \
a1 = __op__(a1, s1); \
a2 = __op__(a2, s2); \
} \
\
dst[0] = (acctype)a0; \
dst[1] = (acctype)a1; \
dst[2] = (acctype)a2; \
} \
\
return CV_OK; \
}
#define ICV_ACC_COLS_FUNC_C4( name, flavor, arrtype, worktype, acctype, __op__ ) \
static CvStatus CV_STDCALL \
icv##name##Cols_##flavor##_C4R( const arrtype* src, int srcstep, \
acctype* dst, int dststep, CvSize size )\
{ \
int i, width = size.width*4; \
srcstep /= sizeof(src[0]); \
dststep /= sizeof(dst[0]); \
\
for( ; size.height--; src += srcstep, dst += dststep ) \
{ \
worktype a0 = src[0], a1 = src[1], a2 = src[2], a3 = src[3]; \
for( i = 4; i < width; i += 4 ) \
{ \
worktype s0 = src[i], s1 = src[i+1]; \
a0 = __op__(a0, s0); \
a1 = __op__(a1, s1); \
s0 = src[i+2]; s1 = src[i+3]; \
a2 = __op__(a2, s0); \
a3 = __op__(a3, s1); \
} \
\
dst[0] = (acctype)a0; \
dst[1] = (acctype)a1; \
dst[2] = (acctype)a2; \
dst[3] = (acctype)a3; \
} \
\
return CV_OK; \
}
ICV_ACC_ROWS_FUNC( Sum, 8u32s, uchar, int, CV_ADD, CV_NOP )
ICV_ACC_ROWS_FUNC( Sum, 8u32f, uchar, float, CV_ADD, CV_8TO32F )
ICV_ACC_ROWS_FUNC( Sum, 16u32f, ushort, float, CV_ADD, CV_NOP )
ICV_ACC_ROWS_FUNC( Sum, 16u64f, ushort, double, CV_ADD, CV_NOP )
ICV_ACC_ROWS_FUNC( Sum, 16s32f, short, float, CV_ADD, CV_NOP )
ICV_ACC_ROWS_FUNC( Sum, 16s64f, short, double, CV_ADD, CV_NOP )
ICV_ACC_ROWS_FUNC( Sum, 32f, float, float, CV_ADD, CV_NOP )
ICV_ACC_ROWS_FUNC( Sum, 32f64f, float, double, CV_ADD, CV_NOP )
ICV_ACC_ROWS_FUNC( Sum, 64f, double, double, CV_ADD, CV_NOP )
ICV_ACC_ROWS_FUNC( Max, 8u, uchar, uchar, CV_MAX_8U, CV_NOP )
ICV_ACC_ROWS_FUNC( Max, 32f, float, float, MAX, CV_NOP )
ICV_ACC_ROWS_FUNC( Max, 64f, double, double, MAX, CV_NOP )
ICV_ACC_ROWS_FUNC( Min, 8u, uchar, uchar, CV_MIN_8U, CV_NOP )
ICV_ACC_ROWS_FUNC( Min, 32f, float, float, MIN, CV_NOP )
ICV_ACC_ROWS_FUNC( Min, 64f, double, double, MIN, CV_NOP )
ICV_ACC_COLS_FUNC_C1( Sum, 8u32s, uchar, int, int, CV_ADD )
ICV_ACC_COLS_FUNC_C1( Sum, 8u32f, uchar, int, float, CV_ADD )
ICV_ACC_COLS_FUNC_C1( Sum, 16u32f, ushort, float, float, CV_ADD )
ICV_ACC_COLS_FUNC_C1( Sum, 16u64f, ushort, double, double, CV_ADD )
ICV_ACC_COLS_FUNC_C1( Sum, 16s32f, short, float, float, CV_ADD )
ICV_ACC_COLS_FUNC_C1( Sum, 16s64f, short, double, double, CV_ADD )
ICV_ACC_COLS_FUNC_C1( Sum, 32f, float, float, float, CV_ADD )
ICV_ACC_COLS_FUNC_C1( Sum, 32f64f, float, double, double, CV_ADD )
ICV_ACC_COLS_FUNC_C1( Sum, 64f, double, double, double, CV_ADD )
ICV_ACC_COLS_FUNC_C3( Sum, 8u32s, uchar, int, int, CV_ADD )
ICV_ACC_COLS_FUNC_C3( Sum, 8u32f, uchar, int, float, CV_ADD )
ICV_ACC_COLS_FUNC_C3( Sum, 32f, float, float, float, CV_ADD )
ICV_ACC_COLS_FUNC_C3( Sum, 64f, double, double, double, CV_ADD )
ICV_ACC_COLS_FUNC_C4( Sum, 8u32s, uchar, int, int, CV_ADD )
ICV_ACC_COLS_FUNC_C4( Sum, 8u32f, uchar, int, float, CV_ADD )
ICV_ACC_COLS_FUNC_C4( Sum, 32f, float, float, float, CV_ADD )
ICV_ACC_COLS_FUNC_C4( Sum, 64f, double, double, double, CV_ADD )
ICV_ACC_COLS_FUNC_C1( Max, 8u, uchar, int, uchar, CV_MAX_8U )
ICV_ACC_COLS_FUNC_C1( Max, 32f, float, float, float, MAX )
ICV_ACC_COLS_FUNC_C1( Max, 64f, double, double, double, MAX )
ICV_ACC_COLS_FUNC_C1( Min, 8u, uchar, int, uchar, CV_MIN_8U )
ICV_ACC_COLS_FUNC_C1( Min, 32f, float, float, float, MIN )
ICV_ACC_COLS_FUNC_C1( Min, 64f, double, double, double, MIN )
typedef CvStatus (CV_STDCALL * CvReduceToRowFunc)
( const void* src, int srcstep, void* dst, CvSize size );
typedef CvStatus (CV_STDCALL * CvReduceToColFunc)
( const void* src, int srcstep, void* dst, int dststep, CvSize size );
CV_IMPL void
cvReduce( const CvArr* srcarr, CvArr* dstarr, int dim, int op )
{
CvMat* temp = 0;
CV_FUNCNAME( "cvReduce" );
__BEGIN__;
CvMat sstub, *src = (CvMat*)srcarr;
CvMat dstub, *dst = (CvMat*)dstarr, *dst0;
int sdepth, ddepth, cn, op0 = op;
CvSize size;
if( !CV_IS_MAT(src) )
CV_CALL( src = cvGetMat( src, &sstub ));
if( !CV_IS_MAT(dst) )
CV_CALL( dst = cvGetMat( dst, &dstub ));
if( !CV_ARE_CNS_EQ(src, dst) )
CV_ERROR( CV_StsUnmatchedFormats, "Input and output arrays must have the same number of channels" );
sdepth = CV_MAT_DEPTH(src->type);
ddepth = CV_MAT_DEPTH(dst->type);
cn = CV_MAT_CN(src->type);
dst0 = dst;
size = cvGetMatSize(src);
if( dim < 0 )
dim = src->rows > dst->rows ? 0 : src->cols > dst->cols ? 1 : dst->cols == 1;
if( dim > 1 )
CV_ERROR( CV_StsOutOfRange, "The reduced dimensionality index is out of range" );
if( (dim == 0 && (dst->cols != src->cols || dst->rows != 1)) ||
(dim == 1 && (dst->rows != src->rows || dst->cols != 1)) )
CV_ERROR( CV_StsBadSize, "The output array size is incorrect" );
if( op == CV_REDUCE_AVG )
{
int ttype = sdepth == CV_8U ? CV_MAKETYPE(CV_32S,cn) : dst->type;
if( ttype != dst->type )
CV_CALL( dst = temp = cvCreateMat( dst->rows, dst->cols, ttype ));
op = CV_REDUCE_SUM;
ddepth = CV_MAT_DEPTH(ttype);
}
if( op != CV_REDUCE_SUM && op != CV_REDUCE_MAX && op != CV_REDUCE_MIN )
CV_ERROR( CV_StsBadArg, "Unknown reduce operation index, must be one of CV_REDUCE_*" );
if( dim == 0 )
{
CvReduceToRowFunc rfunc =
op == CV_REDUCE_SUM ?
(sdepth == CV_8U && ddepth == CV_32S ? (CvReduceToRowFunc)icvSumRows_8u32s_C1R :
sdepth == CV_8U && ddepth == CV_32F ? (CvReduceToRowFunc)icvSumRows_8u32f_C1R :
sdepth == CV_16U && ddepth == CV_32F ? (CvReduceToRowFunc)icvSumRows_16u32f_C1R :
sdepth == CV_16U && ddepth == CV_64F ? (CvReduceToRowFunc)icvSumRows_16u64f_C1R :
sdepth == CV_16S && ddepth == CV_32F ? (CvReduceToRowFunc)icvSumRows_16s32f_C1R :
sdepth == CV_16S && ddepth == CV_64F ? (CvReduceToRowFunc)icvSumRows_16s64f_C1R :
sdepth == CV_32F && ddepth == CV_32F ? (CvReduceToRowFunc)icvSumRows_32f_C1R :
sdepth == CV_32F && ddepth == CV_64F ? (CvReduceToRowFunc)icvSumRows_32f64f_C1R :
sdepth == CV_64F && ddepth == CV_64F ? (CvReduceToRowFunc)icvSumRows_64f_C1R : 0) :
op == CV_REDUCE_MAX ?
(sdepth == CV_8U && ddepth == CV_8U ? (CvReduceToRowFunc)icvMaxRows_8u_C1R :
sdepth == CV_32F && ddepth == CV_32F ? (CvReduceToRowFunc)icvMaxRows_32f_C1R :
sdepth == CV_64F && ddepth == CV_64F ? (CvReduceToRowFunc)icvMaxRows_64f_C1R : 0) :
(sdepth == CV_8U && ddepth == CV_8U ? (CvReduceToRowFunc)icvMinRows_8u_C1R :
sdepth == CV_32F && ddepth == CV_32F ? (CvReduceToRowFunc)icvMinRows_32f_C1R :
sdepth == CV_64F && ddepth == CV_64F ? (CvReduceToRowFunc)icvMinRows_64f_C1R : 0);
if( !rfunc )
CV_ERROR( CV_StsUnsupportedFormat,
"Unsupported combination of input and output array formats" );
size.width *= cn;
IPPI_CALL( rfunc( src->data.ptr, src->step ? src->step : CV_STUB_STEP,
dst->data.ptr, size ));
}
else
{
CvReduceToColFunc cfunc =
op == CV_REDUCE_SUM ?
(sdepth == CV_8U && ddepth == CV_32S ?
(CvReduceToColFunc)(cn == 1 ? icvSumCols_8u32s_C1R :
cn == 3 ? icvSumCols_8u32s_C3R :
cn == 4 ? icvSumCols_8u32s_C4R : 0) :
sdepth == CV_8U && ddepth == CV_32F ?
(CvReduceToColFunc)(cn == 1 ? icvSumCols_8u32f_C1R :
cn == 3 ? icvSumCols_8u32f_C3R :
cn == 4 ? icvSumCols_8u32f_C4R : 0) :
sdepth == CV_16U && ddepth == CV_32F ?
(CvReduceToColFunc)(cn == 1 ? icvSumCols_16u32f_C1R : 0) :
sdepth == CV_16U && ddepth == CV_64F ?
(CvReduceToColFunc)(cn == 1 ? icvSumCols_16u64f_C1R : 0) :
sdepth == CV_16S && ddepth == CV_32F ?
(CvReduceToColFunc)(cn == 1 ? icvSumCols_16s32f_C1R : 0) :
sdepth == CV_16S && ddepth == CV_64F ?
(CvReduceToColFunc)(cn == 1 ? icvSumCols_16s64f_C1R : 0) :
sdepth == CV_32F && ddepth == CV_32F ?
(CvReduceToColFunc)(cn == 1 ? icvSumCols_32f_C1R :
cn == 3 ? icvSumCols_32f_C3R :
cn == 4 ? icvSumCols_32f_C4R : 0) :
sdepth == CV_32F && ddepth == CV_64F ?
(CvReduceToColFunc)(cn == 1 ? icvSumCols_32f64f_C1R : 0) :
sdepth == CV_64F && ddepth == CV_64F ?
(CvReduceToColFunc)(cn == 1 ? icvSumCols_64f_C1R :
cn == 3 ? icvSumCols_64f_C3R :
cn == 4 ? icvSumCols_64f_C4R : 0) : 0) :
op == CV_REDUCE_MAX && cn == 1 ?
(sdepth == CV_8U && ddepth == CV_8U ? (CvReduceToColFunc)icvMaxCols_8u_C1R :
sdepth == CV_32F && ddepth == CV_32F ? (CvReduceToColFunc)icvMaxCols_32f_C1R :
sdepth == CV_64F && ddepth == CV_64F ? (CvReduceToColFunc)icvMaxCols_64f_C1R : 0) :
op == CV_REDUCE_MIN && cn == 1 ?
(sdepth == CV_8U && ddepth == CV_8U ? (CvReduceToColFunc)icvMinCols_8u_C1R :
sdepth == CV_32F && ddepth == CV_32F ? (CvReduceToColFunc)icvMinCols_32f_C1R :
sdepth == CV_64F && ddepth == CV_64F ? (CvReduceToColFunc)icvMinCols_64f_C1R : 0) : 0;
if( !cfunc )
CV_ERROR( CV_StsUnsupportedFormat,
"Unsupported combination of input and output array formats" );
IPPI_CALL( cfunc( src->data.ptr, src->step ? src->step : CV_STUB_STEP,
dst->data.ptr, dst->step ? dst->step : CV_STUB_STEP, size ));
}
if( op0 == CV_REDUCE_AVG )
cvScale( dst, dst0, 1./(dim == 0 ? src->rows : src->cols) );
__END__;
if( temp )
cvReleaseMat( &temp );
}
/* End of file. */