#include "OptimizeCal.h" const unsigned char icvSaturate8u_cv[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 }; void FastCopy(unsigned char* pSrc, unsigned char* pDst, int nWidth) { //set a _mm256 to save the src Data Width while (nWidth > 0) { __m256i srcData = _mm256_loadu_si256((__m256i*)pSrc); __m256i src2Data = _mm256_loadu_si256((__m256i*)(pSrc + 32)); __m256i* pDstData = (__m256i*)pDst; __m256i* pDst2Data = (__m256i*)(pDst + 32); _mm256_storeu_si256(pDstData, srcData); _mm256_storeu_si256(pDst2Data, src2Data); pSrc += 64; pDst += 64; nWidth -= 64; } } void distanceATS_L1_8u(unsigned char* const src, int nWidth, int nHeight, int nPitch, unsigned char* dst) { int width = nWidth, height = nHeight; int a; unsigned char lut[256]; int x, y; const unsigned char* sbase = src; unsigned char* dbase = dst; int srcstep = (int)nPitch; int dststep = (int)nPitch; ////////////////////// forward scan //////////////////////// for (x = 0; x < 256; x++) lut[x] = cv::saturate_cast(x + 1); //init first pixel to max (we're going to be skipping it) dbase[0] = (unsigned char)(sbase[0] == 0 ? 0 : 255); //first row (scan west only, skip first pixel) for (x = 1; x < width; x++) dbase[x] = (unsigned char)(sbase[x] == 0 ? 0 : lut[dbase[x - 1]]); for (y = 1; y < height; y++) { sbase += srcstep; dbase += dststep; //for left edge, scan north only a = sbase[0] == 0 ? 0 : lut[dbase[-dststep]]; dbase[0] = (unsigned char)a; for (x = 1; x < width; x++) { a = sbase[x] == 0 ? 0 : lut[std::min(a, (int)dbase[x - dststep])]; dbase[x] = (unsigned char)a; } } ////////////////////// backward scan /////////////////////// a = dbase[width - 1]; // do last row east pixel scan here (skip bottom right pixel) for (x = width - 2; x >= 0; x--) { a = lut[a]; dbase[x] = (unsigned char)(CV_CALC_MIN_8U(a, dbase[x])); } // right edge is the only error case for (y = height - 2; y >= 0; y--) { dbase -= dststep; // do right edge a = lut[dbase[width - 1 + dststep]]; a = dbase[width - 1] = (unsigned char)(std::min(a,(int)dbase[width - 1])); for (x = width - 2; x >= 0; x--) { int b = dbase[x + dststep]; a = lut[std::min(a, b)]; a = std::min(a, (int)dbase[x]); dbase[x] = (unsigned char)(a); } } } void FastCopy_Any(const unsigned char* src_ptr, unsigned char* dst_ptr, int width) { SIMD_ALIGNED(unsigned char vin[128]); SIMD_ALIGNED(unsigned char vout[128]); memset(vin, 0, sizeof(vin)); /* for YUY2 and msan */ int r = width & 63; int n = width & ~63; if (n > 0) { FastCopy((unsigned char*)src_ptr, dst_ptr, n); } memcpy(vin, src_ptr + (n >> 0) * 1, SS(r, 0) * 1); FastCopy(vin, vout, 63 + 1); memcpy(dst_ptr + n * 1, vout, r * 1); } void OptimizeCalMatAdd_8U(unsigned char* pRelated, int nRelatedWidth, int nRelatedHeight, int nRelatedPitch, unsigned char* pTarget, int nTargetWidth, int nTargetHeight, int nTargetPitch, int nLeft, int nTop, int nChannel ) { int nCap = 256 / (8 * nChannel); unsigned char* pTargetPart = pTarget + nTop * nTargetPitch + nLeft; __m256i mmRelate,mmTarget; __m256i mmResult; for (int i = 0; i < nRelatedHeight; i++) { for (int wid = 0; wid < nRelatedWidth; wid+= nCap) { if (nRelatedWidth - wid < nCap) { int nlast = nRelatedWidth - wid; memcpy(mmRelate.m256i_u8, pRelated + i * nRelatedPitch + wid * nChannel, nlast * nChannel); memcpy(mmTarget.m256i_u8, pTargetPart + i * nTargetPitch + wid * nChannel, nlast * nChannel); } else { _mm256_loadu_epi8(pTargetPart + i * nTargetPitch + wid * nChannel); _mm256_loadu_epi8(pRelated + i * nRelatedPitch + wid * nChannel); } mmResult = _mm256_adds_epu8(mmRelate, mmTarget); if (nRelatedWidth - wid < nCap) { int nlast = nRelatedWidth - wid; memcpy(pTargetPart + i * nTargetPitch + wid* nChannel, mmResult.m256i_u8, nlast * nChannel); } else _mm256_storeu_epi8(pTargetPart + i * nTargetPitch + wid* nChannel, mmResult); } } }