CV_IPP_CHECK()
{
if (ippiCopy_8u_C1MR(_src, (int)sstep, _dst, (int)dstep, ippiSize(size), mask, (int)mstep) >= 0)
{
CV_IMPL_ADD(CV_IMPL_IPP);
return;
}
setIppErrorStatus();
}
inline int arithm_ipp_not8u(const uchar* src1, size_t step1, uchar* dst, size_t step, int width, int height)
{
if (!CV_IPP_CHECK_COND)
return 0;
if (height == 1)
step1 = step = width * sizeof(dst[0]);
if (0 <= CV_INSTRUMENT_FUN_IPP(ippiNot_8u_C1R, src1, (int)step1, dst, (int)step, ippiSize(width, height)))
{
CV_IMPL_ADD(CV_IMPL_IPP);
return 1;
}
setIppErrorStatus();
return 0;
}
static bool IPPDeriv(const Mat& src, Mat& dst, int ddepth, int dx, int dy, int ksize, double scale)
{
int bufSize = 0;
cv::AutoBuffer<char> buffer;
if (ksize == 3 || ksize == 5)
{
if ( ddepth < 0 )
ddepth = src.depth();
if (src.type() == CV_8U && dst.type() == CV_16S && scale == 1)
{
if ((dx == 1) && (dy == 0))
{
if (0 > ippiFilterSobelNegVertGetBufferSize_8u16s_C1R(ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),&bufSize))
return false;
buffer.allocate(bufSize);
return (0 <= ippiFilterSobelNegVertBorder_8u16s_C1R((const Ipp8u*)src.data, (int)src.step,
(Ipp16s*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer));
}
if ((dx == 0) && (dy == 1))
{
if (0 > ippiFilterSobelHorizGetBufferSize_8u16s_C1R(ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),&bufSize))
return false;
buffer.allocate(bufSize);
return (0 <= ippiFilterSobelHorizBorder_8u16s_C1R((const Ipp8u*)src.data, (int)src.step,
(Ipp16s*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer));
}
if ((dx == 2) && (dy == 0))
{
if (0 > ippiFilterSobelVertSecondGetBufferSize_8u16s_C1R(ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),&bufSize))
return false;
buffer.allocate(bufSize);
return (0 <= ippiFilterSobelVertSecondBorder_8u16s_C1R((const Ipp8u*)src.data, (int)src.step,
(Ipp16s*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer));
}
if ((dx == 0) && (dy == 2))
{
if (0 > ippiFilterSobelHorizSecondGetBufferSize_8u16s_C1R(ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),&bufSize))
return false;
buffer.allocate(bufSize);
return (0 <= ippiFilterSobelHorizSecondBorder_8u16s_C1R((const Ipp8u*)src.data, (int)src.step,
(Ipp16s*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer));
}
}
if (src.type() == CV_32F && dst.type() == CV_32F)
{
#if defined(HAVE_IPP_ICV_ONLY) // N/A: ippiMulC_32f_C1R
return false;
#else
#if 0
if ((dx == 1) && (dy == 0))
{
if (0 > ippiFilterSobelNegVertGetBufferSize_32f_C1R(ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize), &bufSize))
return false;
buffer.allocate(bufSize);
if (0 > ippiFilterSobelNegVertBorder_32f_C1R((const Ipp32f*)src.data, (int)src.step,
(Ipp32f*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer))
{
return false;
}
if(scale != 1)
ippiMulC_32f_C1R((Ipp32f *)dst.data, (int)dst.step, (Ipp32f)scale, (Ipp32f *)dst.data, (int)dst.step, ippiSize(dst.cols*dst.channels(), dst.rows));
return true;
}
if ((dx == 0) && (dy == 1))
{
if (0 > ippiFilterSobelHorizGetBufferSize_32f_C1R(ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),&bufSize))
return false;
buffer.allocate(bufSize);
if (0 > ippiFilterSobelHorizBorder_32f_C1R((const Ipp32f*)src.data, (int)src.step,
(Ipp32f*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer))
{
return false;
}
if(scale != 1)
ippiMulC_32f_C1R((Ipp32f *)dst.data, (int)dst.step, (Ipp32f)scale, (Ipp32f *)dst.data, (int)dst.step, ippiSize(dst.cols*dst.channels(), dst.rows));
return true;
}
#endif
if((dx == 2) && (dy == 0))
{
if (0 > ippiFilterSobelVertSecondGetBufferSize_32f_C1R(ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),&bufSize))
return false;
buffer.allocate(bufSize);
if (0 > ippiFilterSobelVertSecondBorder_32f_C1R((const Ipp32f*)src.data, (int)src.step,
(Ipp32f*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer))
{
return false;
}
if(scale != 1)
ippiMulC_32f_C1R((Ipp32f *)dst.data, (int)dst.step, (Ipp32f)scale, (Ipp32f *)dst.data, (int)dst.step, ippiSize(dst.cols*dst.channels(), dst.rows));
return true;
}
if((dx == 0) && (dy == 2))
{
if (0 > ippiFilterSobelHorizSecondGetBufferSize_32f_C1R(ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),&bufSize))
return false;
buffer.allocate(bufSize);
if (0 > ippiFilterSobelHorizSecondBorder_32f_C1R((const Ipp32f*)src.data, (int)src.step,
(Ipp32f*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows), (IppiMaskSize)(ksize*10+ksize),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer))
{
return false;
}
if(scale != 1)
ippiMulC_32f_C1R((Ipp32f *)dst.data, (int)dst.step, (Ipp32f)scale, (Ipp32f *)dst.data, (int)dst.step, ippiSize(dst.cols*dst.channels(), dst.rows));
return true;
}
#endif
}
}
if(ksize <= 0)
return IPPDerivScharr(src, dst, ddepth, dx, dy, scale);
return false;
}
static bool IPPDerivScharr(const Mat& src, Mat& dst, int ddepth, int dx, int dy, double scale)
{
int bufSize = 0;
cv::AutoBuffer<char> buffer;
IppiSize roi = ippiSize(src.cols, src.rows);
if( ddepth < 0 )
ddepth = src.depth();
dst.create( src.size(), CV_MAKETYPE(ddepth, src.channels()) );
switch(src.type())
{
case CV_8U:
{
if(scale != 1)
return false;
switch(dst.type())
{
case CV_16S:
{
if ((dx == 1) && (dy == 0))
{
if (0 > ippiFilterScharrVertGetBufferSize_8u16s_C1R(roi,&bufSize))
return false;
buffer.allocate(bufSize);
return (0 <= ippiFilterScharrVertBorder_8u16s_C1R((const Ipp8u*)src.data, (int)src.step,
(Ipp16s*)dst.data, (int)dst.step, roi, ippBorderRepl, 0, (Ipp8u*)(char*)buffer));
}
if ((dx == 0) && (dy == 1))
{
if (0 > ippiFilterScharrHorizGetBufferSize_8u16s_C1R(roi,&bufSize))
return false;
buffer.allocate(bufSize);
return (0 <= ippiFilterScharrHorizBorder_8u16s_C1R((const Ipp8u*)src.data, (int)src.step,
(Ipp16s*)dst.data, (int)dst.step, roi, ippBorderRepl, 0, (Ipp8u*)(char*)buffer));
}
return false;
}
default:
return false;
}
}
case CV_32F:
#if defined(HAVE_IPP_ICV_ONLY) // N/A: ippiMulC_32f_C1R
return false;
#else
{
switch(dst.type())
{
case CV_32F:
{
if ((dx == 1) && (dy == 0))
{
if (0 > ippiFilterScharrVertGetBufferSize_32f_C1R(ippiSize(src.cols, src.rows),&bufSize))
return false;
buffer.allocate(bufSize);
if (0 > ippiFilterScharrVertBorder_32f_C1R((const Ipp32f*)src.data, (int)src.step,
(Ipp32f*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer))
{
return false;
}
if (scale != 1)
/* IPP is fast, so MulC produce very little perf degradation.*/
//ippiMulC_32f_C1IR((Ipp32f)scale, (Ipp32f*)dst.data, (int)dst.step, ippiSize(dst.cols*dst.channels(), dst.rows));
ippiMulC_32f_C1R((Ipp32f*)dst.data, (int)dst.step, (Ipp32f)scale, (Ipp32f*)dst.data, (int)dst.step, ippiSize(dst.cols*dst.channels(), dst.rows));
return true;
}
if ((dx == 0) && (dy == 1))
{
if (0 > ippiFilterScharrHorizGetBufferSize_32f_C1R(ippiSize(src.cols, src.rows),&bufSize))
return false;
buffer.allocate(bufSize);
if (0 > ippiFilterScharrHorizBorder_32f_C1R((const Ipp32f*)src.data, (int)src.step,
(Ipp32f*)dst.data, (int)dst.step, ippiSize(src.cols, src.rows),
ippBorderRepl, 0, (Ipp8u*)(char*)buffer))
return false;
if (scale != 1)
ippiMulC_32f_C1R((Ipp32f *)dst.data, (int)dst.step, (Ipp32f)scale, (Ipp32f *)dst.data, (int)dst.step, ippiSize(dst.cols*dst.channels(), dst.rows));
return true;
}
}
default:
return false;
}
}
#endif
default:
return false;
}
}
inline int arithm_ipp_sub32f(const float* src1, size_t step1, const float* src2, size_t step2,
float* dst, size_t step, int width, int height)
{
ARITHM_IPP_BIN(ippiSub_32f_C1R, src2, (int)step2, src1, (int)step1, dst, (int)step, ippiSize(width, height));
}
inline int arithm_ipp_xor8u(const uchar* src1, size_t step1, const uchar* src2, size_t step2,
uchar* dst, size_t step, int width, int height)
{
ARITHM_IPP_BIN(ippiXor_8u_C1R, src1, (int)step1, src2, (int)step2, dst, (int)step, ippiSize(width, height));
}
inline int arithm_ipp_sub8u(const uchar* src1, size_t step1, const uchar* src2, size_t step2,
uchar* dst, size_t step, int width, int height)
{
ARITHM_IPP_BIN(ippiSub_8u_C1RSfs, src2, (int)step2, src1, (int)step1, dst, (int)step, ippiSize(width, height), 0);
}
inline int arithm_ipp_add16u(const ushort* src1, size_t step1, const ushort* src2, size_t step2,
ushort* dst, size_t step, int width, int height)
{
ARITHM_IPP_BIN(ippiAdd_16u_C1RSfs, src1, (int)step1, src2, (int)step2, dst, (int)step, ippiSize(width, height), 0);
}
inline int arithm_ipp_mul32f(const float *src1, size_t step1, const float *src2, size_t step2,
float *dst, size_t step, int width, int height, double scale)
{
ARITHM_IPP_MUL(ippiMul_32f_C1R, src1, (int)step1, src2, (int)step2,dst, (int)step, ippiSize(width, height));
}
inline int arithm_ipp_mul16s(const short *src1, size_t step1, const short *src2, size_t step2,
short *dst, size_t step, int width, int height, double scale)
{
ARITHM_IPP_MUL(ippiMul_16s_C1RSfs, src1, (int)step1, src2, (int)step2,dst, (int)step, ippiSize(width, height), 0);
}
inline int arithm_ipp_mul8u(const uchar *src1, size_t step1, const uchar *src2, size_t step2,
uchar *dst, size_t step, int width, int height, double scale)
{
ARITHM_IPP_MUL(ippiMul_8u_C1RSfs, src1, (int)step1, src2, (int)step2,dst, (int)step, ippiSize(width, height), 0);
}
inline int arithm_ipp_cmp32f(const float* src1, size_t step1, const float* src2, size_t step2,
uchar* dst, size_t step, int width, int height, int cmpop)
{
ARITHM_IPP_CMP(ippiCompare_32f_C1R, src1, (int)step1, src2, (int)step2, dst, (int)step, ippiSize(width, height));
}
if( mask[x+2] )
dst[x+2] = src[x+2];
if( mask[x+3] )
dst[x+3] = src[x+3];
}
#endif
for( ; x < size.width; x++ )
if( mask[x] )
dst[x] = src[x];
}
}
template<> void
copyMask_<uchar>(const uchar* _src, size_t sstep, const uchar* mask, size_t mstep, uchar* _dst, size_t dstep, Size size)
{
CV_IPP_RUN(true, ippiCopy_8u_C1MR(_src, (int)sstep, _dst, (int)dstep, ippiSize(size), mask, (int)mstep) >= 0)
for( ; size.height--; mask += mstep, _src += sstep, _dst += dstep )
{
const uchar* src = (const uchar*)_src;
uchar* dst = (uchar*)_dst;
int x = 0;
#if CV_SSE4_2
if(USE_SSE4_2)//
{
__m128i zero = _mm_setzero_si128 ();
for( ; x <= size.width - 16; x += 16 )
{
const __m128i rSrc = _mm_lddqu_si128((const __m128i*)(src+x));
__m128i _mask = _mm_lddqu_si128((const __m128i*)(mask+x));
inline int arithm_ipp_sub16s(const short* src1, size_t step1, const short* src2, size_t step2,
short* dst, size_t step, int width, int height)
{
ARITHM_IPP_BIN(ippiSub_16s_C1RSfs, src2, (int)step2, src1, (int)step1, dst, (int)step, ippiSize(width, height), 0);
}
void cv::updateMotionHistory( InputArray _silhouette, InputOutputArray _mhi,
double timestamp, double duration )
{
CV_Assert( _silhouette.type() == CV_8UC1 && _mhi.type() == CV_32FC1 );
CV_Assert( _silhouette.sameSize(_mhi) );
float ts = (float)timestamp;
float delbound = (float)(timestamp - duration);
CV_OCL_RUN(_mhi.isUMat() && _mhi.dims() <= 2,
ocl_updateMotionHistory(_silhouette, _mhi, ts, delbound))
Mat silh = _silhouette.getMat(), mhi = _mhi.getMat();
Size size = silh.size();
#ifdef HAVE_IPP
int silhstep = (int)silh.step, mhistep = (int)mhi.step;
#endif
if( silh.isContinuous() && mhi.isContinuous() )
{
size.width *= size.height;
size.height = 1;
#ifdef HAVE_IPP
silhstep = (int)silh.total();
mhistep = (int)mhi.total() * sizeof(Ipp32f);
#endif
}
#ifdef HAVE_IPP
IppStatus status = ippiUpdateMotionHistory_8u32f_C1IR((const Ipp8u *)silh.data, silhstep, (Ipp32f *)mhi.data, mhistep,
ippiSize(size.width, size.height), (Ipp32f)timestamp, (Ipp32f)duration);
if (status >= 0)
return;
#endif
#if CV_SSE2
volatile bool useSIMD = cv::checkHardwareSupport(CV_CPU_SSE2);
#endif
for(int y = 0; y < size.height; y++ )
{
const uchar* silhData = silh.ptr<uchar>(y);
float* mhiData = mhi.ptr<float>(y);
int x = 0;
#if CV_SSE2
if( useSIMD )
{
__m128 ts4 = _mm_set1_ps(ts), db4 = _mm_set1_ps(delbound);
for( ; x <= size.width - 8; x += 8 )
{
__m128i z = _mm_setzero_si128();
__m128i s = _mm_unpacklo_epi8(_mm_loadl_epi64((const __m128i*)(silhData + x)), z);
__m128 s0 = _mm_cvtepi32_ps(_mm_unpacklo_epi16(s, z)), s1 = _mm_cvtepi32_ps(_mm_unpackhi_epi16(s, z));
__m128 v0 = _mm_loadu_ps(mhiData + x), v1 = _mm_loadu_ps(mhiData + x + 4);
__m128 fz = _mm_setzero_ps();
v0 = _mm_and_ps(v0, _mm_cmpge_ps(v0, db4));
v1 = _mm_and_ps(v1, _mm_cmpge_ps(v1, db4));
__m128 m0 = _mm_and_ps(_mm_xor_ps(v0, ts4), _mm_cmpneq_ps(s0, fz));
__m128 m1 = _mm_and_ps(_mm_xor_ps(v1, ts4), _mm_cmpneq_ps(s1, fz));
v0 = _mm_xor_ps(v0, m0);
v1 = _mm_xor_ps(v1, m1);
_mm_storeu_ps(mhiData + x, v0);
_mm_storeu_ps(mhiData + x + 4, v1);
}
}
#endif
for( ; x < size.width; x++ )
{
float val = mhiData[x];
val = silhData[x] ? ts : val < delbound ? 0 : val;
mhiData[x] = val;
}
}
}
if( mask[x+2] )
dst[x+2] = src[x+2];
if( mask[x+3] )
dst[x+3] = src[x+3];
}
#endif
for( ; x < size.width; x++ )
if( mask[x] )
dst[x] = src[x];
}
}
template<> void
copyMask_<uchar>(const uchar* _src, size_t sstep, const uchar* mask, size_t mstep, uchar* _dst, size_t dstep, Size size)
{
CV_IPP_RUN_FAST(CV_INSTRUMENT_FUN_IPP(ippiCopy_8u_C1MR, _src, (int)sstep, _dst, (int)dstep, ippiSize(size), mask, (int)mstep) >= 0)
for( ; size.height--; mask += mstep, _src += sstep, _dst += dstep )
{
const uchar* src = (const uchar*)_src;
uchar* dst = (uchar*)_dst;
int x = 0;
#if CV_SIMD128
{
v_uint8x16 v_zero = v_setzero_u8();
for( ; x <= size.width - 16; x += 16 )
{
v_uint8x16 v_src = v_load(src + x),
v_dst = v_load(dst + x),
v_nmask = v_load(mask + x) == v_zero;
inline int arithm_ipp_absdiff16u(const ushort* src1, size_t step1, const ushort* src2, size_t step2,
ushort* dst, size_t step, int width, int height)
{
ARITHM_IPP_BIN(ippiAbsDiff_16u_C1R, src1, (int)step1, src2, (int)step2, dst, (int)step, ippiSize(width, height));
}
static inline void ippiGetImage(const cv::Mat &src, ::ipp::IwiImage &dst)
{
::ipp::IwiBorderSize inMemBorder;
if(src.isSubmatrix()) // already have physical border
{
cv::Size origSize;
cv::Point offset;
src.locateROI(origSize, offset);
inMemBorder.left = (IwSize)offset.x;
inMemBorder.top = (IwSize)offset.y;
inMemBorder.right = (IwSize)(origSize.width - src.cols - offset.x);
inMemBorder.bottom = (IwSize)(origSize.height - src.rows - offset.y);
}
dst.Init(ippiSize(src.size()), ippiGetDataType(src.depth()), src.channels(), inMemBorder, (void*)src.ptr(), src.step);
}
