CV_IMPL void
cvFilter2D( const CvArr* _src, CvArr* _dst, const CvMat* _kernel, CvPoint anchor )
{
// below that approximate size OpenCV is faster
const int ipp_lower_limit = 20;
static CvFuncTable filter_tab;
static int inittab = 0;
CvFilterState *state = 0;
float* kernel_data = 0;
int local_alloc = 1;
CvMat* temp = 0;
CV_FUNCNAME( "cvFilter2D" );
__BEGIN__;
CvFilterFunc func = 0;
int coi1 = 0, coi2 = 0;
CvMat srcstub, *src = (CvMat*)_src;
CvMat dststub, *dst = (CvMat*)_dst;
CvSize size;
int type, depth;
int src_step, dst_step;
CvMat kernel_hdr;
const CvMat* kernel = _kernel;
if( !inittab )
{
icvInitFilterTab( &filter_tab );
inittab = 1;
}
CV_CALL( src = cvGetMat( src, &srcstub, &coi1 ));
CV_CALL( dst = cvGetMat( dst, &dststub, &coi2 ));
if( coi1 != 0 || coi2 != 0 )
CV_ERROR( CV_BadCOI, "" );
type = CV_MAT_TYPE( src->type );
if( !CV_ARE_SIZES_EQ( src, dst ))
CV_ERROR( CV_StsUnmatchedSizes, "" );
if( !CV_ARE_TYPES_EQ( src, dst ))
CV_ERROR( CV_StsUnmatchedFormats, "" );
if( !CV_IS_MAT(kernel) ||
(CV_MAT_TYPE(kernel->type) != CV_32F &&
CV_MAT_TYPE(kernel->type) != CV_64F ))
CV_ERROR( CV_StsBadArg, "kernel must be single-channel floating-point matrix" );
if( anchor.x == -1 && anchor.y == -1 )
anchor = cvPoint(kernel->cols/2,kernel->rows/2);
if( (unsigned)anchor.x >= (unsigned)kernel->cols ||
(unsigned)anchor.y >= (unsigned)kernel->rows )
CV_ERROR( CV_StsOutOfRange, "anchor point is out of kernel" );
if( CV_MAT_TYPE(kernel->type) != CV_32FC1 || !CV_IS_MAT_CONT(kernel->type) || icvFilter_8u_C1R_p )
{
int sz = kernel->rows*kernel->cols*sizeof(kernel_data[0]);
if( sz < CV_MAX_LOCAL_SIZE )
kernel_data = (float*)cvStackAlloc( sz );
else
{
CV_CALL( kernel_data = (float*)cvAlloc( sz ));
local_alloc = 0;
}
kernel_hdr = cvMat( kernel->rows, kernel->cols, CV_32F, kernel_data );
if( CV_MAT_TYPE(kernel->type) == CV_32FC1 )
cvCopy( kernel, &kernel_hdr );
else
cvConvertScale( kernel, &kernel_hdr, 1, 0 );
kernel = &kernel_hdr;
}
size = cvGetMatSize( src );
depth = CV_MAT_DEPTH(type);
src_step = src->step;
dst_step = dst->step ? dst->step : CV_STUB_STEP;
if( icvFilter_8u_C1R_p && (src->rows >= ipp_lower_limit || src->cols >= ipp_lower_limit) )
{
CvFilterIPPFunc ipp_func =
type == CV_8UC1 ? (CvFilterIPPFunc)icvFilter_8u_C1R_p :
type == CV_8UC3 ? (CvFilterIPPFunc)icvFilter_8u_C3R_p :
type == CV_8UC4 ? (CvFilterIPPFunc)icvFilter_8u_C4R_p :
type == CV_16SC1 ? (CvFilterIPPFunc)icvFilter_16s_C1R_p :
type == CV_16SC3 ? (CvFilterIPPFunc)icvFilter_16s_C3R_p :
type == CV_16SC4 ? (CvFilterIPPFunc)icvFilter_16s_C4R_p :
type == CV_32FC1 ? (CvFilterIPPFunc)icvFilter_32f_C1R_p :
type == CV_32FC3 ? (CvFilterIPPFunc)icvFilter_32f_C3R_p :
type == CV_32FC4 ? (CvFilterIPPFunc)icvFilter_32f_C4R_p : 0;
if( ipp_func )
{
CvSize el_size = { kernel->cols, kernel->rows };
CvPoint el_anchor = { el_size.width - anchor.x - 1, el_size.height - anchor.y - 1 };
int stripe_size = 1 << 16; // the optimal value may depend on CPU cache,
// overhead of current IPP code etc.
const uchar* shifted_ptr;
int i, j, y, dy = 0;
int temp_step;
// mirror the kernel around the center
for( i = 0; i < (el_size.height+1)/2; i++ )
{
float* top_row = kernel->data.fl + el_size.width*i;
float* bottom_row = kernel->data.fl + el_size.width*(el_size.height - i - 1);
for( j = 0; j < (el_size.width+1)/2; j++ )
{
float a = top_row[j], b = top_row[el_size.width - j - 1];
float c = bottom_row[j], d = bottom_row[el_size.width - j - 1];
top_row[j] = d;
top_row[el_size.width - j - 1] = c;
bottom_row[j] = b;
bottom_row[el_size.width - j - 1] = a;
}
}
CV_CALL( temp = icvIPPFilterInit( src, stripe_size, el_size ));
shifted_ptr = temp->data.ptr +
anchor.y*temp->step + anchor.x*CV_ELEM_SIZE(type);
temp_step = temp->step ? temp->step : CV_STUB_STEP;
for( y = 0; y < src->rows; y += dy )
{
dy = icvIPPFilterNextStripe( src, temp, y, el_size, anchor );
IPPI_CALL( ipp_func( shifted_ptr, temp_step,
dst->data.ptr + y*dst_step, dst_step, cvSize(src->cols, dy),
kernel->data.fl, el_size, el_anchor ));
}
EXIT;
}
}
CV_CALL( state = icvFilterInitAlloc( src->cols, cv32f, CV_MAT_CN(type),
cvSize(kernel->cols, kernel->rows), anchor,
kernel->data.ptr, ICV_GENERIC_KERNEL ));
if( CV_MAT_CN(type) == 2 )
CV_ERROR( CV_BadNumChannels, "Unsupported number of channels" );
func = (CvFilterFunc)(filter_tab.fn_2d[depth]);
if( !func )
CV_ERROR( CV_StsUnsupportedFormat, "" );
if( size.height == 1 )
src_step = dst_step = CV_STUB_STEP;
IPPI_CALL( func( src->data.ptr, src_step, dst->data.ptr,
dst_step, &size, state, 0 ));
__END__;
cvReleaseMat( &temp );
icvFilterFree( &state );
if( !local_alloc )
cvFree( (void**)&kernel_data );
}
CV_IMPL void
cvMatchTemplate( const CvArr* _img, const CvArr* _templ, CvArr* _result, int method )
{
CvMat* sum = 0;
CvMat* sqsum = 0;
CV_FUNCNAME( "cvMatchTemplate" );
__BEGIN__;
int coi1 = 0, coi2 = 0;
int depth, cn;
int i, j, k;
CvMat stub, *img = (CvMat*)_img;
CvMat tstub, *templ = (CvMat*)_templ;
CvMat rstub, *result = (CvMat*)_result;
CvScalar templ_mean = cvScalarAll(0);
double templ_norm = 0, templ_sum2 = 0;
int idx = 0, idx2 = 0;
double *p0, *p1, *p2, *p3;
double *q0, *q1, *q2, *q3;
double inv_area;
int sum_step, sqsum_step;
int num_type = method == CV_TM_CCORR || method == CV_TM_CCORR_NORMED ? 0 :
method == CV_TM_CCOEFF || method == CV_TM_CCOEFF_NORMED ? 1 : 2;
int is_normed = method == CV_TM_CCORR_NORMED ||
method == CV_TM_SQDIFF_NORMED ||
method == CV_TM_CCOEFF_NORMED;
CV_CALL( img = cvGetMat( img, &stub, &coi1 ));
CV_CALL( templ = cvGetMat( templ, &tstub, &coi2 ));
CV_CALL( result = cvGetMat( result, &rstub ));
if( CV_MAT_DEPTH( img->type ) != CV_8U &&
CV_MAT_DEPTH( img->type ) != CV_32F )
CV_ERROR( CV_StsUnsupportedFormat,
"The function supports only 8u and 32f data types" );
if( !CV_ARE_TYPES_EQ( img, templ ))
CV_ERROR( CV_StsUnmatchedSizes, "image and template should have the same type" );
if( CV_MAT_TYPE( result->type ) != CV_32FC1 )
CV_ERROR( CV_StsUnsupportedFormat, "output image should have 32f type" );
if( img->rows < templ->rows || img->cols < templ->cols )
{
CvMat* t;
CV_SWAP( img, templ, t );
}
if( result->rows != img->rows - templ->rows + 1 ||
result->cols != img->cols - templ->cols + 1 )
CV_ERROR( CV_StsUnmatchedSizes, "output image should be (W - w + 1)x(H - h + 1)" );
if( method < CV_TM_SQDIFF || method > CV_TM_CCOEFF_NORMED )
CV_ERROR( CV_StsBadArg, "unknown comparison method" );
depth = CV_MAT_DEPTH(img->type);
cn = CV_MAT_CN(img->type);
if( is_normed && cn == 1 && templ->rows > 8 && templ->cols > 8 &&
img->rows > templ->cols && img->cols > templ->cols )
{
CvTemplMatchIPPFunc ipp_func =
depth == CV_8U ?
(method == CV_TM_SQDIFF_NORMED ? (CvTemplMatchIPPFunc)icvSqrDistanceValid_Norm_8u32f_C1R_p :
method == CV_TM_CCORR_NORMED ? (CvTemplMatchIPPFunc)icvCrossCorrValid_Norm_8u32f_C1R_p :
(CvTemplMatchIPPFunc)icvCrossCorrValid_NormLevel_8u32f_C1R_p) :
(method == CV_TM_SQDIFF_NORMED ? (CvTemplMatchIPPFunc)icvSqrDistanceValid_Norm_32f_C1R_p :
method == CV_TM_CCORR_NORMED ? (CvTemplMatchIPPFunc)icvCrossCorrValid_Norm_32f_C1R_p :
(CvTemplMatchIPPFunc)icvCrossCorrValid_NormLevel_32f_C1R_p);
if( ipp_func )
{
CvSize img_size = cvGetMatSize(img), templ_size = cvGetMatSize(templ);
IPPI_CALL( ipp_func( img->data.ptr, img->step ? img->step : CV_STUB_STEP,
img_size, templ->data.ptr,
templ->step ? templ->step : CV_STUB_STEP,
templ_size, result->data.ptr,
result->step ? result->step : CV_STUB_STEP ));
for( i = 0; i < result->rows; i++ )
{
float* rrow = (float*)(result->data.ptr + i*result->step);
for( j = 0; j < result->cols; j++ )
{
if( fabs(rrow[j]) > 1. )
rrow[j] = rrow[j] < 0 ? -1.f : 1.f;
}
}
EXIT;
}
}
CV_CALL( icvCrossCorr( img, templ, result ));
if( method == CV_TM_CCORR )
EXIT;
inv_area = 1./((double)templ->rows * templ->cols);
CV_CALL( sum = cvCreateMat( img->rows + 1, img->cols + 1,
CV_MAKETYPE( CV_64F, cn )));
if( method == CV_TM_CCOEFF )
{
CV_CALL( cvIntegral( img, sum, 0, 0 ));
CV_CALL( templ_mean = cvAvg( templ ));
q0 = q1 = q2 = q3 = 0;
}
else
{
CvScalar _templ_sdv = cvScalarAll(0);
CV_CALL( sqsum = cvCreateMat( img->rows + 1, img->cols + 1,
CV_MAKETYPE( CV_64F, cn )));
CV_CALL( cvIntegral( img, sum, sqsum, 0 ));
CV_CALL( cvAvgSdv( templ, &templ_mean, &_templ_sdv ));
templ_norm = CV_SQR(_templ_sdv.val[0]) + CV_SQR(_templ_sdv.val[1]) +
CV_SQR(_templ_sdv.val[2]) + CV_SQR(_templ_sdv.val[3]);
if( templ_norm < DBL_EPSILON && method == CV_TM_CCOEFF_NORMED )
{
cvSet( result, cvScalarAll(1.) );
EXIT;
}
templ_sum2 = templ_norm +
CV_SQR(templ_mean.val[0]) + CV_SQR(templ_mean.val[1]) +
CV_SQR(templ_mean.val[2]) + CV_SQR(templ_mean.val[3]);
if( num_type != 1 )
{
templ_mean = cvScalarAll(0);
templ_norm = templ_sum2;
}
templ_sum2 /= inv_area;
templ_norm = sqrt(templ_norm);
templ_norm /= sqrt(inv_area); // care of accuracy here
q0 = (double*)sqsum->data.ptr;
q1 = q0 + templ->cols*cn;
q2 = (double*)(sqsum->data.ptr + templ->rows*sqsum->step);
q3 = q2 + templ->cols*cn;
}
p0 = (double*)sum->data.ptr;
p1 = p0 + templ->cols*cn;
p2 = (double*)(sum->data.ptr + templ->rows*sum->step);
p3 = p2 + templ->cols*cn;
sum_step = sum ? sum->step / sizeof(double) : 0;
sqsum_step = sqsum ? sqsum->step / sizeof(double) : 0;
for( i = 0; i < result->rows; i++ )
{
float* rrow = (float*)(result->data.ptr + i*result->step);
idx = i * sum_step;
idx2 = i * sqsum_step;
for( j = 0; j < result->cols; j++, idx += cn, idx2 += cn )
{
double num = rrow[j], t;
double wnd_mean2 = 0, wnd_sum2 = 0;
if( num_type == 1 )
{
for( k = 0; k < cn; k++ )
{
t = p0[idx+k] - p1[idx+k] - p2[idx+k] + p3[idx+k];
wnd_mean2 += CV_SQR(t);
num -= t*templ_mean.val[k];
}
wnd_mean2 *= inv_area;
}
if( is_normed || num_type == 2 )
{
for( k = 0; k < cn; k++ )
{
t = q0[idx2+k] - q1[idx2+k] - q2[idx2+k] + q3[idx2+k];
wnd_sum2 += t;
}
if( num_type == 2 )
num = wnd_sum2 - 2*num + templ_sum2;
}
if( is_normed )
{
t = sqrt(MAX(wnd_sum2 - wnd_mean2,0))*templ_norm;
if( t > DBL_EPSILON )
{
num /= t;
if( fabs(num) > 1. )
num = num > 0 ? 1 : -1;
}
else
num = method != CV_TM_SQDIFF_NORMED || num < DBL_EPSILON ? 0 : 1;
}
rrow[j] = (float)num;
}
}
__END__;
cvReleaseMat( &sum );
cvReleaseMat( &sqsum );
}
/* Wrapper function for distance transform group */
CV_IMPL void
cvDistTransform( const void* srcarr, void* dstarr,
int distType, int maskSize,
const float *mask,
void* labelsarr )
{
CvMat* temp = 0;
CvMat* src_copy = 0;
CvMemStorage* st = 0;
CV_FUNCNAME( "cvDistTransform" );
__BEGIN__;
float _mask[5] = {0};
int _imask[3];
CvMat srcstub, *src = (CvMat*)srcarr;
CvMat dststub, *dst = (CvMat*)dstarr;
CvMat lstub, *labels = (CvMat*)labelsarr;
CvSize size;
CvIPPDistTransFunc ipp_func = 0;
CvIPPDistTransFunc2 ipp_inp_func = 0;
CV_CALL( src = cvGetMat( src, &srcstub ));
CV_CALL( dst = cvGetMat( dst, &dststub ));
if( !CV_IS_MASK_ARR( src ) || CV_MAT_TYPE( dst->type ) != CV_32FC1 &&
(CV_MAT_TYPE(dst->type) != CV_8UC1 || distType != CV_DIST_L1 || labels) )
CV_ERROR( CV_StsUnsupportedFormat,
"source image must be 8uC1 and the distance map must be 32fC1 "
"(or 8uC1 in case of simple L1 distance transform)" );
if( !CV_ARE_SIZES_EQ( src, dst ))
CV_ERROR( CV_StsUnmatchedSizes, "the source and the destination images must be of the same size" );
if( maskSize != CV_DIST_MASK_3 && maskSize != CV_DIST_MASK_5 && maskSize != CV_DIST_MASK_PRECISE )
CV_ERROR( CV_StsBadSize, "Mask size should be 3 or 5 or 0 (presize)" );
if( distType == CV_DIST_C || distType == CV_DIST_L1 )
maskSize = !labels ? CV_DIST_MASK_3 : CV_DIST_MASK_5;
else if( distType == CV_DIST_L2 && labels )
maskSize = CV_DIST_MASK_5;
if( maskSize == CV_DIST_MASK_PRECISE )
{
CV_CALL( icvTrueDistTrans( src, dst ));
EXIT;
}
if( labels )
{
CV_CALL( labels = cvGetMat( labels, &lstub ));
if( CV_MAT_TYPE( labels->type ) != CV_32SC1 )
CV_ERROR( CV_StsUnsupportedFormat, "the output array of labels must be 32sC1" );
if( !CV_ARE_SIZES_EQ( labels, dst ))
CV_ERROR( CV_StsUnmatchedSizes, "the array of labels has a different size" );
if( maskSize == CV_DIST_MASK_3 )
CV_ERROR( CV_StsNotImplemented,
"3x3 mask can not be used for \"labeled\" distance transform. Use 5x5 mask" );
}
if( distType == CV_DIST_C || distType == CV_DIST_L1 || distType == CV_DIST_L2 )
{
icvGetDistanceTransformMask( (distType == CV_DIST_C ? 0 :
distType == CV_DIST_L1 ? 1 : 2) + maskSize*10, _mask );
}
else if( distType == CV_DIST_USER )
{
if( !mask )
CV_ERROR( CV_StsNullPtr, "" );
memcpy( _mask, mask, (maskSize/2 + 1)*sizeof(float));
}
if( !labels )
{
if( CV_MAT_TYPE(dst->type) == CV_32FC1 )
ipp_func = (CvIPPDistTransFunc)(maskSize == CV_DIST_MASK_3 ?
icvDistanceTransform_3x3_8u32f_C1R_p : icvDistanceTransform_5x5_8u32f_C1R_p);
else if( src->data.ptr != dst->data.ptr )
ipp_func = (CvIPPDistTransFunc)icvDistanceTransform_3x3_8u_C1R_p;
else
ipp_inp_func = icvDistanceTransform_3x3_8u_C1IR_p;
}
size = cvGetMatSize(src);
if( (ipp_func || ipp_inp_func) && src->cols >= 4 && src->rows >= 2 )
{
_imask[0] = cvRound(_mask[0]);
_imask[1] = cvRound(_mask[1]);
_imask[2] = cvRound(_mask[2]);
if( ipp_func )
{
IPPI_CALL( ipp_func( src->data.ptr, src->step,
dst->data.fl, dst->step, size,
CV_MAT_TYPE(dst->type) == CV_8UC1 ?
(void*)_imask : (void*)_mask ));
}
else
{
IPPI_CALL( ipp_inp_func( src->data.ptr, src->step, size, _imask ));
}
}
else if( CV_MAT_TYPE(dst->type) == CV_8UC1 )
{
CV_CALL( icvDistanceATS_L1_8u( src, dst ));
}
else
{
int border = maskSize == CV_DIST_MASK_3 ? 1 : 2;
CV_CALL( temp = cvCreateMat( size.height + border*2, size.width + border*2, CV_32SC1 ));
if( !labels )
{
CvDistTransFunc func = maskSize == CV_DIST_MASK_3 ?
icvDistanceTransform_3x3_C1R :
icvDistanceTransform_5x5_C1R;
func( src->data.ptr, src->step, temp->data.i, temp->step,
dst->data.fl, dst->step, size, _mask );
}
else
{
CvSeq *contours = 0;
CvPoint top_left = {0,0}, bottom_right = {size.width-1,size.height-1};
int label;
CV_CALL( st = cvCreateMemStorage() );
CV_CALL( src_copy = cvCreateMat( size.height, size.width, src->type ));
cvCmpS( src, 0, src_copy, CV_CMP_EQ );
cvFindContours( src_copy, st, &contours, sizeof(CvContour),
CV_RETR_CCOMP, CV_CHAIN_APPROX_SIMPLE );
cvZero( labels );
for( label = 1; contours != 0; contours = contours->h_next, label++ )
{
CvScalar area_color = cvScalarAll(label);
cvDrawContours( labels, contours, area_color, area_color, -255, -1, 8 );
}
cvCopy( src, src_copy );
cvRectangle( src_copy, top_left, bottom_right, cvScalarAll(255), 1, 8 );
icvDistanceTransformEx_5x5_C1R( src_copy->data.ptr, src_copy->step, temp->data.i, temp->step,
dst->data.fl, dst->step, labels->data.i, labels->step, size, _mask );
}
}
__END__;
cvReleaseMat( &temp );
cvReleaseMat( &src_copy );
cvReleaseMemStorage( &st );
}
