Add threaded version of equalizeHist

This commit is contained in:
Daniil Osokin 2012-12-18 16:10:11 +04:00 committed by Andrey Kamaev
parent 0bbba847a4
commit 98d7d99244

View File

@ -2404,6 +2404,146 @@ cvCalcProbDensity( const CvHistogram* hist, const CvHistogram* hist_mask,
}
}
class EqualizeHistCalcHist_Invoker
{
public:
enum {HIST_SZ = 256};
#ifdef HAVE_TBB
typedef tbb::mutex* MutextPtr;
#else
typedef void* MutextPtr;
#endif
EqualizeHistCalcHist_Invoker(cv::Mat& src, int* histogram, MutextPtr histogramLock)
: src_(src), globalHistogram_(histogram), histogramLock_(histogramLock)
{ }
void operator()( const cv::BlockedRange& rowRange ) const
{
int localHistogram[HIST_SZ] = {0, };
const size_t sstep = src_.step;
int width = src_.cols;
int height = rowRange.end() - rowRange.begin();
if (src_.isContinuous())
{
width *= height;
height = 1;
}
for (const uchar* ptr = src_.ptr<uchar>(rowRange.begin()); height--; ptr += sstep)
{
int x = 0;
for (; x <= width - 4; x += 4)
{
int t0 = ptr[x], t1 = ptr[x+1];
localHistogram[t0]++; localHistogram[t1]++;
t0 = ptr[x+2]; t1 = ptr[x+3];
localHistogram[t0]++; localHistogram[t1]++;
}
for (; x < width; ++x, ++ptr)
localHistogram[ptr[x]]++;
}
#ifdef HAVE_TBB
tbb::mutex::scoped_lock lock(*histogramLock_);
#endif
for( int i = 0; i < HIST_SZ; i++ )
globalHistogram_[i] += localHistogram[i];
}
static bool isWorthParallel( const cv::Mat& src )
{
#ifdef HAVE_TBB
return ( src.total() >= 640*480 );
#else
(void)src;
return false;
#endif
}
private:
EqualizeHistCalcHist_Invoker& operator=(const EqualizeHistCalcHist_Invoker&);
cv::Mat& src_;
int* globalHistogram_;
MutextPtr histogramLock_;
};
class EqualizeHistLut_Invoker
{
public:
EqualizeHistLut_Invoker( cv::Mat& src, cv::Mat& dst, int* lut )
: src_(src),
dst_(dst),
lut_(lut)
{ }
void operator()( const cv::BlockedRange& rowRange ) const
{
const size_t sstep = src_.step;
const size_t dstep = dst_.step;
int width = src_.cols;
int height = rowRange.end() - rowRange.begin();
int* lut = lut_;
if (src_.isContinuous() && dst_.isContinuous())
{
width *= height;
height = 1;
}
const uchar* sptr = src_.ptr<uchar>(rowRange.begin());
uchar* dptr = dst_.ptr<uchar>(rowRange.begin());
for (; height--; sptr += sstep, dptr += dstep)
{
int x = 0;
for (; x <= width - 4; x += 4)
{
int v0 = sptr[x];
int v1 = sptr[x+1];
int x0 = lut[v0];
int x1 = lut[v1];
dptr[x] = (uchar)x0;
dptr[x+1] = (uchar)x1;
v0 = sptr[x+2];
v1 = sptr[x+3];
x0 = lut[v0];
x1 = lut[v1];
dptr[x+2] = (uchar)x0;
dptr[x+3] = (uchar)x1;
}
for (; x < width; ++x)
dptr[x] = (uchar)lut[sptr[x]];
}
}
static bool isWorthParallel( const cv::Mat& src )
{
#ifdef HAVE_TBB
return ( src.total() >= 640*480 );
#else
(void)src;
return false;
#endif
}
private:
EqualizeHistLut_Invoker& operator=(const EqualizeHistLut_Invoker&);
cv::Mat& src_;
cv::Mat& dst_;
int* lut_;
};
CV_IMPL void cvEqualizeHist( const CvArr* srcarr, CvArr* dstarr )
{
@ -2421,35 +2561,25 @@ void cv::equalizeHist( InputArray _src, OutputArray _dst )
if(src.empty())
return;
const int hist_sz = (1 << (8*sizeof(uchar)));
#ifdef HAVE_TBB
tbb::mutex histogramLockInstance;
EqualizeHistCalcHist_Invoker::MutextPtr histogramLock = &histogramLockInstance;
#else
EqualizeHistCalcHist_Invoker::MutextPtr histogramLock = 0;
#endif
const int hist_sz = EqualizeHistCalcHist_Invoker::HIST_SZ;
int hist[hist_sz] = {0,};
int lut[hist_sz];
const size_t sstep = src.step;
const size_t dstep = dst.step;
EqualizeHistCalcHist_Invoker calcBody(src, hist, histogramLock);
EqualizeHistLut_Invoker lutBody(src, dst, lut);
cv::BlockedRange heightRange(0, src.rows);
int width = src.cols;
int height = src.rows;
if (src.isContinuous())
{
width *= height;
height = 1;
}
for (const uchar* ptr = src.ptr<uchar>(); height--; ptr += sstep)
{
int x = 0;
for (; x <= width - 4; x += 4)
{
int t0 = ptr[x], t1 = ptr[x+1];
hist[t0]++; hist[t1]++;
t0 = ptr[x+2]; t1 = ptr[x+3];
hist[t0]++; hist[t1]++;
}
for (; x < width; ++x, ++ptr)
hist[ptr[x]]++;
}
if(EqualizeHistCalcHist_Invoker::isWorthParallel(src))
parallel_for(heightRange, calcBody);
else
calcBody(heightRange);
int i = 0;
while (!hist[i]) ++i;
@ -2464,49 +2594,16 @@ void cv::equalizeHist( InputArray _src, OutputArray _dst )
float scale = (hist_sz - 1.f)/(total - hist[i]);
int sum = 0;
int lut[hist_sz];
for (lut[i++] = 0; i < hist_sz; ++i)
{
sum += hist[i];
lut[i] = saturate_cast<uchar>(sum * scale);
}
int cols = src.cols;
int rows = src.rows;
if (src.isContinuous() && dst.isContinuous())
{
cols *= rows;
rows = 1;
}
const uchar* sptr = src.ptr<uchar>();
uchar* dptr = dst.ptr<uchar>();
for (; rows--; sptr += sstep, dptr += dstep)
{
int x = 0;
for (; x <= cols - 4; x += 4)
{
int v0 = sptr[x];
int v1 = sptr[x+1];
int x0 = lut[v0];
int x1 = lut[v1];
dptr[x] = (uchar)x0;
dptr[x+1] = (uchar)x1;
v0 = sptr[x+2];
v1 = sptr[x+3];
x0 = lut[v0];
x1 = lut[v1];
dptr[x+2] = (uchar)x0;
dptr[x+3] = (uchar)x1;
}
for (; x < cols; ++x)
dptr[x] = (uchar)lut[sptr[x]];
}
if(EqualizeHistLut_Invoker::isWorthParallel(src))
parallel_for(heightRange, lutBody);
else
lutBody(heightRange);
}
/* Implementation of RTTI and Generic Functions for CvHistogram */