Skip to content

O
opencv
Commit 8bb987e4 authored by Vadim Pisarevsky's avatar Vadim Pisarevsky
Browse files
Options

optimized VNG demosaicing using SSE2 (got ~3x speedup)

parent 2fe75ea3
Hide whitespace changes
Inline Side-by-side
Showing with 474 additions and 183 deletions
modules/imgproc/src/color.cpp
View file @ 8bb987e4
......@@ -2191,236 +2191,527 @@ icvBayer2BGR_VNG_8u_C1C3R( const uchar* bayer, int bstep,
{
int blueIdx = code == CV_BayerBG2BGR_VNG || code == CV_BayerGB2BGR_VNG ? 0 : 2;
bool greenCell0 = code != CV_BayerBG2BGR_VNG && code != CV_BayerRG2BGR_VNG;
bool greenCell0_buf = !greenCell0;
// for too small images use the simple interpolation algorithm
if( MIN(size.width, size.height) < 5 )
if( MIN(size.width, size.height) < 8 )
return icvBayer2BGR_8u_C1C3R( bayer, bstep, dst, dststep, size, code );
const int brows = 3, bcn = 7;
int i, bufstep = size.width*bcn;
int N = size.width, N2 = N*2, N3 = N*3, N4 = N*4, N5 = N*5, N6 = N*6, N7 = N*7;
int i, bufstep = N7*bcn;
cv::AutoBuffer<ushort> _buf(bufstep*brows);
ushort* buf = (ushort*)_buf;
bayer += bstep*2;
for( i = 0; i < size.width*3; i++ )
{
dst[i] = dst[i + dststep] =
dst[i + dststep*(size.height-2)] =
dst[i + dststep*(size.height-1)] = 0;
}
#if CV_SSE2
bool haveSSE = cv::checkHardwareSupport(CV_CPU_SSE2);
#define _mm_absdiff_epu16(a,b) _mm_adds_epu16(_mm_subs_epu16(a, b), _mm_subs_epu16(b, a))
#endif
for( int y = 2; y < size.height - 2; y++ )
for( int y = 2; y < size.height - 4; y++ )
{
uchar* dstrow = dst + dststep*y;
uchar* dstrow = dst + dststep*y + 6;
const uchar* srow;
for( int dy = (y == 2 ? -1 : 1); dy <= 1; dy++ )
{
ushort* brow = buf + ((y + dy - 1)%brows)*bufstep;
srow = bayer + dy*bstep;
ushort* brow = buf + ((y + dy - 1)%brows)*bufstep + 1;
srow = bayer + (y+dy)*bstep + 1;
for( i = 0; i < bcn; i++ )
brow[i] = brow[i + (size.width-1)*bcn] = 0;
brow[N*i-1] = brow[(N-2) + N*i] = 0;
bool greenCell = greenCell0_buf;
for( i = 1; i < size.width-1; i++ )
i = 1;
#if CV_SSE2
if( haveSSE )
{
brow += bcn;
brow[0] = (ushort)( abs(srow[i-1-bstep] - srow[i-1+bstep]) +
abs(srow[i-bstep] - srow[i+bstep])*2 +
abs(srow[i+1-bstep] - srow[i+1+bstep]));
brow[1] = (ushort)( abs(srow[i-1-bstep] - srow[i+1-bstep]) +
abs(srow[i-1] - srow[i+1])*2 +
abs(srow[i-1+bstep] - srow[i+1+bstep]));
brow[2] = (ushort)(abs(srow[i+1-bstep] - srow[i-1+bstep])*2);
brow[3] = (ushort)(abs(srow[i-1-bstep] - srow[i+1+bstep])*2);
if(!greenCell)
__m128i z = _mm_setzero_si128();
for( ; i <= N-9; i += 8, srow += 8, brow += 8 )
{
brow[4] = (ushort)(brow[2] + abs(srow[i-bstep] - srow[i-1]) + abs(srow[i+bstep] - srow[i+1]));
brow[5] = (ushort)(brow[3] + abs(srow[i-bstep] - srow[i+1]) + abs(srow[i+bstep] - srow[i-1]));
brow[6] = (ushort)((srow[i-bstep] + srow[i-1] + srow[i+1] + srow[i+bstep] + 2)>>2);
__m128i s1, s2, s3, s4, s6, s7, s8, s9;
s1 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-1-bstep)),z);
s2 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep)),z);
s3 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+1-bstep)),z);
s4 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-1)),z);
s6 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+1)),z);
s7 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-1+bstep)),z);
s8 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep)),z);
s9 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+1+bstep)),z);
__m128i b0, b1, b2, b3, b4, b5, b6;
b0 = _mm_adds_epu16(_mm_slli_epi16(_mm_absdiff_epu16(s2,s8),1),
_mm_adds_epu16(_mm_absdiff_epu16(s1, s7),
_mm_absdiff_epu16(s3, s9)));
b1 = _mm_adds_epu16(_mm_slli_epi16(_mm_absdiff_epu16(s4,s6),1),
_mm_adds_epu16(_mm_absdiff_epu16(s1, s3),
_mm_absdiff_epu16(s7, s9)));
b2 = _mm_slli_epi16(_mm_absdiff_epu16(s3,s7),1);
b3 = _mm_slli_epi16(_mm_absdiff_epu16(s1,s9),1);
_mm_storeu_si128((__m128i*)brow, b0);
_mm_storeu_si128((__m128i*)(brow + N), b1);
_mm_storeu_si128((__m128i*)(brow + N2), b2);
_mm_storeu_si128((__m128i*)(brow + N3), b3);
b4 = _mm_adds_epu16(b2,_mm_adds_epu16(_mm_absdiff_epu16(s2, s4),
_mm_absdiff_epu16(s6, s8)));
b5 = _mm_adds_epu16(b3,_mm_adds_epu16(_mm_absdiff_epu16(s2, s6),
_mm_absdiff_epu16(s4, s8)));
b6 = _mm_adds_epu16(_mm_adds_epu16(s2, s4), _mm_adds_epu16(s6, s8));
b6 = _mm_srli_epi16(b6, 1);
_mm_storeu_si128((__m128i*)(brow + N4), b4);
_mm_storeu_si128((__m128i*)(brow + N5), b5);
_mm_storeu_si128((__m128i*)(brow + N6), b6);
}
greenCell = !greenCell;
}
greenCell0_buf = !greenCell0_buf;
#endif
for( ; i < N-1; i++, srow++, brow++ )
{
brow[0] = (ushort)(abs(srow[-1-bstep] - srow[-1+bstep]) +
abs(srow[-bstep] - srow[+bstep])*2 +
abs(srow[1-bstep] - srow[1+bstep]));
brow[N] = (ushort)(abs(srow[-1-bstep] - srow[1-bstep]) +
abs(srow[-1] - srow[1])*2 +
abs(srow[-1+bstep] - srow[1+bstep]));
brow[N2] = (ushort)(abs(srow[+1-bstep] - srow[-1+bstep])*2);
brow[N3] = (ushort)(abs(srow[-1-bstep] - srow[1+bstep])*2);
brow[N4] = (ushort)(brow[N2] + abs(srow[-bstep] - srow[-1]) +
abs(srow[+bstep] - srow[1]));
brow[N5] = (ushort)(brow[N3] + abs(srow[-bstep] - srow[1]) +
abs(srow[+bstep] - srow[-1]));
brow[N6] = (ushort)((srow[-bstep] + srow[-1] + srow[1] + srow[+bstep])>>1);
}
}
const ushort* brow0 = buf + ((y - 2) % brows)*bufstep + bcn;
const ushort* brow1 = buf + ((y - 1) % brows)*bufstep + bcn;
const ushort* brow2 = buf + (y % brows)*bufstep + bcn;
static const float scale[] = { 0.f, 1.f, 0.5f, 0.3333333333f, 0.25f, 0.2f, 0.1666666667f, 0.1428571f, 0.125f };
srow = bayer + y*bstep;
const ushort* brow0 = buf + ((y - 2) % brows)*bufstep + 2;
const ushort* brow1 = buf + ((y - 1) % brows)*bufstep + 2;
const ushort* brow2 = buf + (y % brows)*bufstep + 2;
static const float scale[] = { 0.f, 0.5f, 0.25f, 0.1666666666667f, 0.125f, 0.1f, 0.08333333333f, 0.0714286f, 0.0625f };
srow = bayer + y*bstep + 2;
bool greenCell = greenCell0;
for( i = 0; i < 6; i++ )
dstrow[i] = dstrow[i + (size.width-2)*3] = 0;
dstrow += 6;
i = 2;
#if CV_SSE2
int limit = !haveSSE ? N-2 : greenCell ? std::min(3, N-2) : 2;
#else
int limit = N - 2;
#endif
for( i = 2; i < size.width-2; i++, brow0 += bcn, brow1 += bcn, brow2 += bcn, dstrow += 3 )
do
{
int gradN = (brow0[0] + brow1[0])>>1;
int gradS = (brow1[0] + brow2[0])>>1;
int gradW = (brow1[-bcn+1] + brow1[1])>>1;
int gradE = (brow1[1] + brow1[bcn+1])>>1;
int minGrad = std::min(std::min(std::min(gradN, gradS), gradW), gradE);
int maxGrad = std::max(std::max(std::max(gradN, gradS), gradW), gradE);
int R, G, B;
if( !greenCell )
for( ; i < limit; i++, srow++, brow0++, brow1++, brow2++, dstrow += 3 )
{
int gradNE = (brow0[bcn+4] + brow1[4])>>1;
int gradSW = (brow1[4] + brow2[-bcn+4])>>1;
int gradNW = (brow0[-bcn+5] + brow1[5])>>1;
int gradSE = (brow1[5] + brow2[bcn+5])>>1;
minGrad = std::min(std::min(std::min(std::min(minGrad, gradNE), gradSW), gradNW), gradSE);
maxGrad = std::max(std::max(std::max(std::max(maxGrad, gradNE), gradSW), gradNW), gradSE);
int T = minGrad + maxGrad/2;
int gradN = brow0[0] + brow1[0];
int gradS = brow1[0] + brow2[0];
int gradW = brow1[N-1] + brow1[N];
int gradE = brow1[N] + brow1[N+1];
int minGrad = std::min(std::min(std::min(gradN, gradS), gradW), gradE);
int maxGrad = std::max(std::max(std::max(gradN, gradS), gradW), gradE);
int R, G, B;
int Rs = 0, Gs = 0, Bs = 0, ng = 0;
if( gradN < T )
{
Rs += (srow[i-bstep*2] + srow[i])>>1;
Gs += srow[i-bstep];
Bs += (srow[i-bstep-1] + srow[i-bstep+1])>>1;
ng++;
}
if( gradS < T )
{
Rs += (srow[i+bstep*2] + srow[i])>>1;
Gs += srow[i+bstep];
Bs += (srow[i+bstep-1] + srow[i+bstep+1])>>1;
ng++;
}
if( gradW < T )
{
Rs += (srow[i-2] + srow[i])>>1;
Gs += srow[i-1];
Bs += (srow[i-bstep-1] + srow[i+bstep-1])>>1;
ng++;
}
if( gradE < T )
{
Rs += (srow[i+2] + srow[i])>>1;
Gs += srow[i+1];
Bs += (srow[i-bstep+1] + srow[i+bstep+1])>>1;
ng++;
}
if( gradNE < T )
{
Rs += (srow[i-bstep*2+2] + srow[i])>>1;
Gs += brow0[bcn+6];
Bs += srow[i-bstep+1];
ng++;
}
if( gradSW < T )
{
Rs += (srow[i+bstep*2-2] + srow[i])>>1;
Gs += brow2[-bcn+6];
Bs += srow[i+bstep-1];
ng++;
}
if( gradNW < T )
if( !greenCell )
{
Rs += (srow[i-bstep*2+2] + srow[i])>>1;
Gs += brow0[bcn+6];
Bs += srow[i-bstep+1];
ng++;
int gradNE = brow0[N4+1] + brow1[N4];
int gradSW = brow1[N4] + brow2[N4-1];
int gradNW = brow0[N5-1] + brow1[N5];
int gradSE = brow1[N5] + brow2[N5+1];
minGrad = std::min(std::min(std::min(std::min(minGrad, gradNE), gradSW), gradNW), gradSE);
maxGrad = std::max(std::max(std::max(std::max(maxGrad, gradNE), gradSW), gradNW), gradSE);
int T = minGrad + maxGrad/2;
int Rs = 0, Gs = 0, Bs = 0, ng = 0;
if( gradN < T )
{
Rs += srow[-bstep*2] + srow[0];
Gs += srow[-bstep]*2;
Bs += srow[-bstep-1] + srow[-bstep+1];
ng++;
}
if( gradS < T )
{
Rs += srow[bstep*2] + srow[0];
Gs += srow[bstep]*2;
Bs += srow[bstep-1] + srow[bstep+1];
ng++;
}
if( gradW < T )
{
Rs += srow[-2] + srow[0];
Gs += srow[-1]*2;
Bs += srow[-bstep-1] + srow[bstep-1];
ng++;
}
if( gradE < T )
{
Rs += srow[2] + srow[0];
Gs += srow[1]*2;
Bs += srow[-bstep+1] + srow[bstep+1];
ng++;
}
if( gradNE < T )
{
Rs += srow[-bstep*2+2] + srow[0];
Gs += brow0[N6+1];
Bs += srow[-bstep+1]*2;
ng++;
}
if( gradSW < T )
{
Rs += srow[bstep*2-2] + srow[0];
Gs += brow2[N6-1];
Bs += srow[bstep-1]*2;
ng++;
}
if( gradNW < T )
{
Rs += srow[-bstep*2-2] + srow[0];
Gs += brow0[N6-1];
Bs += srow[-bstep+1]*2;
ng++;
}
if( gradSE < T )
{
Rs += srow[bstep*2+2] + srow[0];
Gs += brow2[N6+1];
Bs += srow[-bstep+1]*2;
ng++;
}
R = srow[0];
G = R + cvRound((Gs - Rs)*scale[ng]);
B = R + cvRound((Bs - Rs)*scale[ng]);
}
if( gradSE < T )
else
{
Rs += (srow[i-bstep*2+2] + srow[i])>>1;
Gs += brow0[bcn+6];
Bs += srow[i-bstep+1];
ng++;
int gradNE = brow0[N2] + brow0[N2+1] + brow1[N2] + brow1[N2+1];
int gradSW = brow1[N2] + brow1[N2-1] + brow2[N2] + brow2[N2-1];
int gradNW = brow0[N3] + brow0[N3-1] + brow1[N3] + brow1[N3-1];
int gradSE = brow1[N3] + brow1[N3+1] + brow2[N3] + brow2[N3+1];
minGrad = std::min(std::min(std::min(std::min(minGrad, gradNE), gradSW), gradNW), gradSE);
maxGrad = std::max(std::max(std::max(std::max(maxGrad, gradNE), gradSW), gradNW), gradSE);
int T = minGrad + maxGrad/2;
int Rs = 0, Gs = 0, Bs = 0, ng = 0;
if( gradN < T )
{
Rs += srow[-bstep*2-1] + srow[-bstep*2+1];
Gs += srow[-bstep*2] + srow[0];
Bs += srow[-bstep]*2;
ng++;
}
if( gradS < T )
{
Rs += srow[bstep*2-1] + srow[bstep*2+1];
Gs += srow[bstep*2] + srow[0];
Bs += srow[bstep]*2;
ng++;
}
if( gradW < T )
{
Rs += srow[-1]*2;
Gs += srow[-2] + srow[0];
Bs += srow[-bstep-2]+srow[bstep-2];
ng++;
}
if( gradE < T )
{
Rs += srow[1]*2;
Gs += srow[2] + srow[0];
Bs += srow[-bstep+2]+srow[bstep+2];
ng++;
}
if( gradNE < T )
{
Rs += srow[-bstep*2+1] + srow[1];
Gs += srow[-bstep+1]*2;
Bs += srow[-bstep] + srow[-bstep+2];
ng++;
}
if( gradSW < T )
{
Rs += srow[bstep*2-1] + srow[-1];
Gs += srow[bstep-1]*2;
Bs += srow[bstep] + srow[bstep-2];
ng++;
}
if( gradNW < T )
{
Rs += srow[-bstep*2-1] + srow[-1];
Gs += srow[-bstep-1]*2;
Bs += srow[-bstep-2]+srow[-bstep];
ng++;
}
if( gradSE < T )
{
Rs += srow[bstep*2+1] + srow[1];
Gs += srow[bstep+1]*2;
Bs += srow[bstep+2]+srow[bstep];
ng++;
}
G = srow[0];
R = G + cvRound((Rs - Gs)*scale[ng]);
B = G + cvRound((Bs - Gs)*scale[ng]);
}
R = srow[i];
G = R + cvRound((Gs - Rs)*scale[ng]);
B = R + cvRound((Bs - Rs)*scale[ng]);
dstrow[blueIdx] = CV_CAST_8U(B);
dstrow[1] = CV_CAST_8U(G);
dstrow[blueIdx^2] = CV_CAST_8U(R);
greenCell = !greenCell;
}
else
#if CV_SSE2
if( !haveSSE )
break;
__m128i emask = _mm_set1_epi32(0x0000ffff),
omask = _mm_set1_epi32(0xffff0000),
z = _mm_setzero_si128();
__m128 _0_5 = _mm_set1_ps(0.5f);
#define _mm_merge_epi16(a, b) \
_mm_or_si128(_mm_and_si128(a, emask), _mm_and_si128(b, omask))
#define _mm_cvtloepi16_ps(a) _mm_cvtepi32_ps(_mm_srai_epi32(_mm_unpacklo_epi16(a,a), 16))
#define _mm_cvthiepi16_ps(a) _mm_cvtepi32_ps(_mm_srai_epi32(_mm_unpackhi_epi16(a,a), 16))
// process 8 pixels at once
for( ; i <= N - 10; i += 8, srow += 8, brow0 += 8, brow1 += 8, brow2 += 8 )
{
int gradNE = (brow0[2] + brow0[bcn+2] + brow1[2] + brow1[bcn+2])>>1;
int gradSW = (brow1[2] + brow1[-bcn+2] + brow2[2] + brow2[-bcn+2])>>1;
int gradNW = (brow0[3] + brow0[-bcn+3] + brow1[3] + brow1[-bcn+3])>>1;
int gradSE = (brow1[3] + brow1[bcn+3] + brow2[3] + brow2[bcn+3])>>1;
__m128i gradN, gradS, gradW, gradE, gradNE, gradSW, gradNW, gradSE;
gradN = _mm_adds_epu16(_mm_loadu_si128((__m128i*)brow0),
_mm_loadu_si128((__m128i*)brow1));
gradS = _mm_adds_epu16(_mm_loadu_si128((__m128i*)brow1),
_mm_loadu_si128((__m128i*)brow2));
gradW = _mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow1+N-1)),
_mm_loadu_si128((__m128i*)(brow1+N)));
gradE = _mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow1+N+1)),
_mm_loadu_si128((__m128i*)(brow1+N)));
minGrad = std::min(std::min(std::min(std::min(minGrad, gradNE), gradSW), gradNW), gradSE);
maxGrad = std::max(std::max(std::max(std::max(maxGrad, gradNE), gradSW), gradNW), gradSE);
int T = minGrad + maxGrad/2;
__m128i minGrad, maxGrad, T;
minGrad = _mm_min_epi16(_mm_min_epi16(_mm_min_epi16(gradN, gradS), gradW), gradE);
maxGrad = _mm_max_epi16(_mm_max_epi16(_mm_max_epi16(gradN, gradS), gradW), gradE);
int Rs = 0, Gs = 0, Bs = 0, ng = 0;
if( gradN < T )
{
Rs += (srow[i-bstep*2-1] + srow[i-bstep*2+1])>>1;
Gs += (srow[i-bstep*2] + srow[i])>>1;
Bs += srow[i-bstep];
ng++;
}
if( gradS < T )
{
Rs += (srow[i+bstep*2-1] + srow[i+bstep*2+1])>>1;
Gs += (srow[i+bstep*2] + srow[i])>>1;
Bs += srow[i+bstep];
ng++;
}
if( gradW < T )
{
Rs += srow[i-1];
Gs += (srow[i-2] + srow[i])>>1;
Bs += (srow[i-bstep-2]+srow[i+bstep-2])>>1;
ng++;
}
if( gradE < T )
{
Rs += srow[i+1];
Gs += (srow[i+2] + srow[i])>>1;
Bs += (srow[i-bstep+2]+srow[i+bstep+2])>>1;
ng++;
}
if( gradNE < T )
{
Rs += (srow[i-bstep*2+1] + srow[i+1])>>1;
Gs += srow[i-bstep+1];
Bs += (srow[i-bstep] + srow[i-bstep+2])>>1;
ng++;
}
if( gradSW < T )
{
Rs += (srow[i+bstep*2-1] + srow[i-1])>>1;
Gs += srow[i+bstep-1];
Bs += (srow[i+bstep] + srow[i+bstep-2])>>1;
ng++;
}
if( gradNW < T )
{
Rs += (srow[i-bstep*2-1] + srow[i-1])>>1;
Gs += srow[i-bstep-1];
Bs += (srow[i-bstep-2]+srow[i-bstep])>>1;
ng++;
}
if( gradSE < T )
__m128i grad0, grad1;
grad0 = _mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow0+N4+1)),
_mm_loadu_si128((__m128i*)(brow1+N4)));
grad1 = _mm_adds_epu16(_mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow0+N2)),
_mm_loadu_si128((__m128i*)(brow0+N2+1))),
_mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow1+N2)),
_mm_loadu_si128((__m128i*)(brow1+N2+1))));
gradNE = _mm_srli_epi16(_mm_merge_epi16(grad0, grad1), 1);
grad0 = _mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow2+N4-1)),
_mm_loadu_si128((__m128i*)(brow1+N4)));
grad1 = _mm_adds_epu16(_mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow2+N2)),
_mm_loadu_si128((__m128i*)(brow2+N2-1))),
_mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow1+N2)),
_mm_loadu_si128((__m128i*)(brow1+N2-1))));
gradSW = _mm_srli_epi16(_mm_merge_epi16(grad0, grad1), 1);
minGrad = _mm_min_epi16(_mm_min_epi16(minGrad, gradNE), gradSW);
maxGrad = _mm_max_epi16(_mm_max_epi16(maxGrad, gradNE), gradSW);
grad0 = _mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow0+N5-1)),
_mm_loadu_si128((__m128i*)(brow1+N5)));
grad1 = _mm_adds_epu16(_mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow0+N3)),
_mm_loadu_si128((__m128i*)(brow0+N3-1))),
_mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow1+N3)),
_mm_loadu_si128((__m128i*)(brow1+N3-1))));
gradNW = _mm_srli_epi16(_mm_merge_epi16(grad0, grad1), 1);
grad0 = _mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow2+N5+1)),
_mm_loadu_si128((__m128i*)(brow1+N5)));
grad1 = _mm_adds_epu16(_mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow2+N3)),
_mm_loadu_si128((__m128i*)(brow2+N3+1))),
_mm_adds_epu16(_mm_loadu_si128((__m128i*)(brow1+N3)),
_mm_loadu_si128((__m128i*)(brow1+N3+1))));
gradSE = _mm_srli_epi16(_mm_merge_epi16(grad0, grad1), 1);
minGrad = _mm_min_epi16(_mm_min_epi16(minGrad, gradNW), gradSE);
maxGrad = _mm_max_epi16(_mm_max_epi16(maxGrad, gradNW), gradSE);
T = _mm_add_epi16(_mm_srli_epi16(maxGrad, 1), minGrad);
__m128i RGs = z, GRs = z, Bs = z, ng = z, mask;
__m128i t0, t1, x0, x1, x2, x3, x4, x5, x6, x7, x8,
x9, x10, x11, x12, x13, x14, x15, x16;
x0 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)srow), z);
x1 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep-1)), z);
x2 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep*2-1)), z);
x3 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep)), z);
x4 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep*2+1)), z);
x5 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep+1)), z);
x6 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep+2)), z);
x7 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+1)), z);
x8 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep+2)), z);
x9 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep+1)), z);
x10 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep*2+1)), z);
x11 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep)), z);
x12 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep*2-1)), z);
x13 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep-1)), z);
x14 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep-2)), z);
x15 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-1)), z);
x16 = _mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep-2)), z);
// gradN
mask = _mm_cmpgt_epi16(T, gradN);
ng = _mm_sub_epi16(ng, mask);
t0 = _mm_slli_epi16(x3, 1);
t1 = _mm_adds_epu16(_mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep*2)), z), x0);
RGs = _mm_adds_epu16(RGs, _mm_and_si128(t1, mask));
GRs = _mm_adds_epu16(GRs, _mm_and_si128(_mm_merge_epi16(t0, _mm_adds_epu16(x2,x4)), mask));
Bs = _mm_adds_epu16(Bs, _mm_and_si128(_mm_merge_epi16(_mm_adds_epu16(x1,x5), t0), mask));
// gradNE
mask = _mm_cmpgt_epi16(T, gradNE);
ng = _mm_sub_epi16(ng, mask);
t0 = _mm_slli_epi16(x5, 1);
t1 = _mm_adds_epu16(_mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep*2+2)), z), x0);
RGs = _mm_adds_epu16(RGs, _mm_and_si128(_mm_merge_epi16(t1, t0), mask));
GRs = _mm_adds_epu16(GRs, _mm_and_si128(_mm_merge_epi16(_mm_loadu_si128((__m128i*)(brow0+N6+1)),
_mm_adds_epu16(x4,x7)), mask));
Bs = _mm_adds_epu16(Bs, _mm_and_si128(_mm_merge_epi16(t0,_mm_adds_epu16(x3,x6)), mask));
// gradE
mask = _mm_cmpgt_epi16(T, gradE);
ng = _mm_sub_epi16(ng, mask);
t0 = _mm_slli_epi16(x7, 1);
t1 = _mm_adds_epu16(_mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+2)), z), x0);
RGs = _mm_adds_epu16(RGs, _mm_and_si128(t1, mask));
GRs = _mm_adds_epu16(GRs, _mm_and_si128(t0, mask));
Bs = _mm_adds_epu16(Bs, _mm_and_si128(_mm_merge_epi16(_mm_adds_epu16(x5,x9),
_mm_adds_epu16(x6,x8)), mask));
// gradSE
mask = _mm_cmpgt_epi16(T, gradSE);
ng = _mm_sub_epi16(ng, mask);
t0 = _mm_slli_epi16(x9, 1);
t1 = _mm_adds_epu16(_mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep*2+2)), z), x0);
RGs = _mm_adds_epu16(RGs, _mm_and_si128(_mm_merge_epi16(t1, t0), mask));
GRs = _mm_adds_epu16(GRs, _mm_and_si128(_mm_merge_epi16(_mm_loadu_si128((__m128i*)(brow2+N6+1)),
_mm_adds_epu16(x7,x10)), mask));
Bs = _mm_adds_epu16(Bs, _mm_and_si128(_mm_merge_epi16(t0, _mm_adds_epu16(x8,x11)), mask));
// gradS
mask = _mm_cmpgt_epi16(T, gradS);
ng = _mm_sub_epi16(ng, mask);
t0 = _mm_slli_epi16(x11, 1);
t1 = _mm_adds_epu16(_mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep*2)), z), x0);
RGs = _mm_adds_epu16(RGs, _mm_and_si128(t1, mask));
GRs = _mm_adds_epu16(GRs, _mm_and_si128(_mm_merge_epi16(t0, _mm_adds_epu16(x10,x12)), mask));
Bs = _mm_adds_epu16(Bs, _mm_and_si128(_mm_merge_epi16(_mm_adds_epu16(x9,x13), t0), mask));
// gradSW
mask = _mm_cmpgt_epi16(T, gradSW);
ng = _mm_sub_epi16(ng, mask);
t0 = _mm_slli_epi16(x13, 1);
t1 = _mm_adds_epu16(_mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow+bstep*2-2)), z), x0);
RGs = _mm_adds_epu16(RGs, _mm_and_si128(_mm_merge_epi16(t1, t0), mask));
GRs = _mm_adds_epu16(GRs, _mm_and_si128(_mm_merge_epi16(_mm_loadu_si128((__m128i*)(brow2+N6-1)),
_mm_adds_epu16(x12,x15)), mask));
Bs = _mm_adds_epu16(Bs, _mm_and_si128(_mm_merge_epi16(t0,_mm_adds_epu16(x11,x14)), mask));
// gradW
mask = _mm_cmpgt_epi16(T, gradW);
ng = _mm_sub_epi16(ng, mask);
t0 = _mm_slli_epi16(x15, 1);
t1 = _mm_adds_epu16(_mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-2)), z), x0);
RGs = _mm_adds_epu16(RGs, _mm_and_si128(t1, mask));
GRs = _mm_adds_epu16(GRs, _mm_and_si128(t0, mask));
Bs = _mm_adds_epu16(Bs, _mm_and_si128(_mm_merge_epi16(_mm_adds_epu16(x1,x13),
_mm_adds_epu16(x14,x16)), mask));
// gradNW
mask = _mm_cmpgt_epi16(T, gradNW);
ng = _mm_sub_epi16(ng, mask);
__m128 ngf0, ngf1;
ngf0 = _mm_div_ps(_0_5, _mm_cvtloepi16_ps(ng));
ngf1 = _mm_div_ps(_0_5, _mm_cvthiepi16_ps(ng));
t0 = _mm_slli_epi16(x1, 1);
t1 = _mm_adds_epu16(_mm_unpacklo_epi8(_mm_loadl_epi64((__m128i*)(srow-bstep*2-2)), z), x0);
RGs = _mm_adds_epu16(RGs, _mm_and_si128(_mm_merge_epi16(t1, t0), mask));
GRs = _mm_adds_epu16(GRs, _mm_and_si128(_mm_merge_epi16(_mm_loadu_si128((__m128i*)(brow0+N6-1)),
_mm_adds_epu16(x2,x15)), mask));
Bs = _mm_adds_epu16(Bs, _mm_and_si128(_mm_merge_epi16(t0,_mm_adds_epu16(x3,x16)), mask));
// now interpolate r, g & b
t0 = _mm_sub_epi16(GRs, RGs);
t1 = _mm_sub_epi16(Bs, RGs);
t0 = _mm_add_epi16(x0, _mm_packs_epi32(
_mm_cvtps_epi32(_mm_mul_ps(_mm_cvtloepi16_ps(t0), ngf0)),
_mm_cvtps_epi32(_mm_mul_ps(_mm_cvthiepi16_ps(t0), ngf1))));
t1 = _mm_add_epi16(x0, _mm_packs_epi32(
_mm_cvtps_epi32(_mm_mul_ps(_mm_cvtloepi16_ps(t1), ngf0)),
_mm_cvtps_epi32(_mm_mul_ps(_mm_cvthiepi16_ps(t1), ngf1))));
x1 = _mm_merge_epi16(x0, t0);
x2 = _mm_merge_epi16(t0, x0);
uchar R[8], G[8], B[8];
_mm_storel_epi64(blueIdx ? (__m128i*)B : (__m128i*)R, _mm_packus_epi16(x1, z));
_mm_storel_epi64((__m128i*)G, _mm_packus_epi16(x2, z));
_mm_storel_epi64(blueIdx ? (__m128i*)R : (__m128i*)B, _mm_packus_epi16(t1, z));
for( int j = 0; j < 8; j++, dstrow += 3 )
{
Rs += (srow[i+bstep*2+1] + srow[i+1])>>1;
Gs += srow[i+bstep+1];
Bs += (srow[i+bstep+2]+srow[i+bstep])>>1;
ng++;
dstrow[0] = B[j]; dstrow[1] = G[j]; dstrow[2] = R[j];
}
G = srow[i];
R = G + cvRound((Rs - Gs)*scale[ng]);
B = G + cvRound((Bs - Gs)*scale[ng]);
}
dstrow[blueIdx] = CV_CAST_8U(B);
dstrow[1] = CV_CAST_8U(G);
dstrow[blueIdx^2] = CV_CAST_8U(R);
greenCell = !greenCell;
#endif
limit = N - 2;
}
while( i < N - 2 );
for( i = 0; i < 6; i++ )
{
dst[dststep*y + 5 - i] = dst[dststep*y + 8 - i];
dst[dststep*y + (N - 2)*3 + i] = dst[dststep*y + (N - 3)*3 + i];
}
greenCell0 = !greenCell0;
blueIdx ^= 2;
}
for( i = 0; i < size.width*3; i++ )
{
dst[i] = dst[i + dststep] = dst[i + dststep*2];
dst[i + dststep*(size.height-4)] =
dst[i + dststep*(size.height-3)] =
dst[i + dststep*(size.height-2)] =
dst[i + dststep*(size.height-1)] = dst[i + dststep*(size.height-5)];
}
return CV_OK;
}
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment