/*M/////////////////////////////////////////////////////////////////////////////////////// // // IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING. // // By downloading, copying, installing or using the software you agree to this license. // If you do not agree to this license, do not download, install, // copy or use the software. // // // Intel License Agreement // For Open Source Computer Vision Library // // Copyright (C) 2000, Intel Corporation, all rights reserved. // Third party copyrights are property of their respective owners. // // Redistribution and use in source and binary forms, with or without modification, // are permitted provided that the following conditions are met: // // * Redistribution's of source code must retain the above copyright notice, // this list of conditions and the following disclaimer. // // * Redistribution's in binary form must reproduce the above copyright notice, // this list of conditions and the following disclaimer in the documentation // and/or other materials provided with the distribution. // // * The name of Intel Corporation may not be used to endorse or promote products // derived from this software without specific prior written permission. // // This software is provided by the copyright holders and contributors "as is" and // any express or implied warranties, including, but not limited to, the implied // warranties of merchantability and fitness for a particular purpose are disclaimed. // In no event shall the Intel Corporation or contributors be liable for any direct, // indirect, incidental, special, exemplary, or consequential damages // (including, but not limited to, procurement of substitute goods or services; // loss of use, data, or profits; or business interruption) however caused // and on any theory of liability, whether in contract, strict liability, // or tort (including negligence or otherwise) arising in any way out of // the use of this software, even if advised of the possibility of such damage. // //M*/ #include "precomp.hpp" #include "utils.hpp" namespace cv { int validateToInt(size_t sz) { int valueInt = (int)sz; CV_Assert((size_t)valueInt == sz); return valueInt; } #define SCALE 14 #define cR (int)(0.299*(1 << SCALE) + 0.5) #define cG (int)(0.587*(1 << SCALE) + 0.5) #define cB ((1 << SCALE) - cR - cG) void icvCvt_BGR2Gray_8u_C3C1R( const uchar* bgr, int bgr_step, uchar* gray, int gray_step, Size size, int _swap_rb ) { int i; for( ; size.height--; gray += gray_step ) { short cBGR0 = cB; short cBGR2 = cR; if (_swap_rb) std::swap(cBGR0, cBGR2); for( i = 0; i < size.width; i++, bgr += 3 ) { int t = descale( bgr[0]*cBGR0 + bgr[1]*cG + bgr[2]*cBGR2, SCALE ); gray[i] = (uchar)t; } bgr += bgr_step - size.width*3; } } void icvCvt_BGRA2Gray_16u_CnC1R( const ushort* bgr, int bgr_step, ushort* gray, int gray_step, Size size, int ncn, int _swap_rb ) { int i; for( ; size.height--; gray += gray_step ) { short cBGR0 = cB; short cBGR2 = cR; if (_swap_rb) std::swap(cBGR0, cBGR2); for( i = 0; i < size.width; i++, bgr += ncn ) { int t = descale( bgr[0]*cBGR0 + bgr[1]*cG + bgr[2]*cBGR2, SCALE ); gray[i] = (ushort)t; } bgr += bgr_step - size.width*ncn; } } void icvCvt_BGRA2Gray_8u_C4C1R( const uchar* bgra, int rgba_step, uchar* gray, int gray_step, Size size, int _swap_rb ) { int i; for( ; size.height--; gray += gray_step ) { short cBGR0 = cB; short cBGR2 = cR; if (_swap_rb) std::swap(cBGR0, cBGR2); for( i = 0; i < size.width; i++, bgra += 4 ) { int t = descale( bgra[0]*cBGR0 + bgra[1]*cG + bgra[2]*cBGR2, SCALE ); gray[i] = (uchar)t; } bgra += rgba_step - size.width*4; } } void icvCvt_Gray2BGR_8u_C1C3R( const uchar* gray, int gray_step, uchar* bgr, int bgr_step, Size size ) { int i; for( ; size.height--; gray += gray_step ) { for( i = 0; i < size.width; i++, bgr += 3 ) { bgr[0] = bgr[1] = bgr[2] = gray[i]; } bgr += bgr_step - size.width*3; } } void icvCvt_Gray2BGR_16u_C1C3R( const ushort* gray, int gray_step, ushort* bgr, int bgr_step, Size size ) { int i; for( ; size.height--; gray += gray_step/sizeof(gray[0]) ) { for( i = 0; i < size.width; i++, bgr += 3 ) { bgr[0] = bgr[1] = bgr[2] = gray[i]; } bgr += bgr_step/sizeof(bgr[0]) - size.width*3; } } void icvCvt_BGRA2BGR_8u_C4C3R( const uchar* bgra, int bgra_step, uchar* bgr, int bgr_step, Size size, int _swap_rb ) { int i; int swap_rb = _swap_rb ? 2 : 0; for( ; size.height--; ) { for( i = 0; i < size.width; i++, bgr += 3, bgra += 4 ) { uchar t0 = bgra[swap_rb], t1 = bgra[1]; bgr[0] = t0; bgr[1] = t1; t0 = bgra[swap_rb^2]; bgr[2] = t0; } bgr += bgr_step - size.width*3; bgra += bgra_step - size.width*4; } } void icvCvt_BGRA2BGR_16u_C4C3R( const ushort* bgra, int bgra_step, ushort* bgr, int bgr_step, Size size, int _swap_rb ) { int i; int swap_rb = _swap_rb ? 2 : 0; for( ; size.height--; ) { for( i = 0; i < size.width; i++, bgr += 3, bgra += 4 ) { ushort t0 = bgra[swap_rb], t1 = bgra[1]; bgr[0] = t0; bgr[1] = t1; t0 = bgra[swap_rb^2]; bgr[2] = t0; } bgr += bgr_step/sizeof(bgr[0]) - size.width*3; bgra += bgra_step/sizeof(bgra[0]) - size.width*4; } } void icvCvt_BGRA2RGBA_8u_C4R( const uchar* bgra, int bgra_step, uchar* rgba, int rgba_step, Size size ) { int i; for( ; size.height--; ) { for( i = 0; i < size.width; i++, bgra += 4, rgba += 4 ) { uchar t0 = bgra[0], t1 = bgra[1]; uchar t2 = bgra[2], t3 = bgra[3]; rgba[0] = t2; rgba[1] = t1; rgba[2] = t0; rgba[3] = t3; } bgra += bgra_step - size.width*4; rgba += rgba_step - size.width*4; } } void icvCvt_BGRA2RGBA_16u_C4R( const ushort* bgra, int bgra_step, ushort* rgba, int rgba_step, Size size ) { int i; for( ; size.height--; ) { for( i = 0; i < size.width; i++, bgra += 4, rgba += 4 ) { ushort t0 = bgra[0], t1 = bgra[1]; ushort t2 = bgra[2], t3 = bgra[3]; rgba[0] = t2; rgba[1] = t1; rgba[2] = t0; rgba[3] = t3; } bgra += bgra_step/sizeof(bgra[0]) - size.width*4; rgba += rgba_step/sizeof(rgba[0]) - size.width*4; } } void icvCvt_BGR2RGB_8u_C3R( const uchar* bgr, int bgr_step, uchar* rgb, int rgb_step, Size size ) { int i; for( ; size.height--; ) { for( i = 0; i < size.width; i++, bgr += 3, rgb += 3 ) { uchar t0 = bgr[0], t1 = bgr[1], t2 = bgr[2]; rgb[2] = t0; rgb[1] = t1; rgb[0] = t2; } bgr += bgr_step - size.width*3; rgb += rgb_step - size.width*3; } } void icvCvt_BGR2RGB_16u_C3R( const ushort* bgr, int bgr_step, ushort* rgb, int rgb_step, Size size ) { int i; for( ; size.height--; ) { for( i = 0; i < size.width; i++, bgr += 3, rgb += 3 ) { ushort t0 = bgr[0], t1 = bgr[1], t2 = bgr[2]; rgb[2] = t0; rgb[1] = t1; rgb[0] = t2; } bgr += bgr_step - size.width*3; rgb += rgb_step - size.width*3; } } typedef unsigned short ushort; void icvCvt_BGR5552Gray_8u_C2C1R( const uchar* bgr555, int bgr555_step, uchar* gray, int gray_step, Size size ) { int i; for( ; size.height--; gray += gray_step, bgr555 += bgr555_step ) { for( i = 0; i < size.width; i++ ) { int t = descale( ((((ushort*)bgr555)[i] << 3) & 0xf8)*cB + ((((ushort*)bgr555)[i] >> 2) & 0xf8)*cG + ((((ushort*)bgr555)[i] >> 7) & 0xf8)*cR, SCALE ); gray[i] = (uchar)t; } } } void icvCvt_BGR5652Gray_8u_C2C1R( const uchar* bgr565, int bgr565_step, uchar* gray, int gray_step, Size size ) { int i; for( ; size.height--; gray += gray_step, bgr565 += bgr565_step ) { for( i = 0; i < size.width; i++ ) { int t = descale( ((((ushort*)bgr565)[i] << 3) & 0xf8)*cB + ((((ushort*)bgr565)[i] >> 3) & 0xfc)*cG + ((((ushort*)bgr565)[i] >> 8) & 0xf8)*cR, SCALE ); gray[i] = (uchar)t; } } } void icvCvt_BGR5552BGR_8u_C2C3R( const uchar* bgr555, int bgr555_step, uchar* bgr, int bgr_step, Size size ) { int i; for( ; size.height--; bgr555 += bgr555_step ) { for( i = 0; i < size.width; i++, bgr += 3 ) { int t0 = (((ushort*)bgr555)[i] << 3) & 0xf8; int t1 = (((ushort*)bgr555)[i] >> 2) & 0xf8; int t2 = (((ushort*)bgr555)[i] >> 7) & 0xf8; bgr[0] = (uchar)t0; bgr[1] = (uchar)t1; bgr[2] = (uchar)t2; } bgr += bgr_step - size.width*3; } } void icvCvt_BGR5652BGR_8u_C2C3R( const uchar* bgr565, int bgr565_step, uchar* bgr, int bgr_step, Size size ) { int i; for( ; size.height--; bgr565 += bgr565_step ) { for( i = 0; i < size.width; i++, bgr += 3 ) { int t0 = (((ushort*)bgr565)[i] << 3) & 0xf8; int t1 = (((ushort*)bgr565)[i] >> 3) & 0xfc; int t2 = (((ushort*)bgr565)[i] >> 8) & 0xf8; bgr[0] = (uchar)t0; bgr[1] = (uchar)t1; bgr[2] = (uchar)t2; } bgr += bgr_step - size.width*3; } } void icvCvt_CMYK2BGR_8u_C4C3R( const uchar* cmyk, int cmyk_step, uchar* bgr, int bgr_step, Size size ) { int i; for( ; size.height--; ) { for( i = 0; i < size.width; i++, bgr += 3, cmyk += 4 ) { int c = cmyk[0], m = cmyk[1], y = cmyk[2], k = cmyk[3]; c = k - ((255 - c)*k>>8); m = k - ((255 - m)*k>>8); y = k - ((255 - y)*k>>8); bgr[2] = (uchar)c; bgr[1] = (uchar)m; bgr[0] = (uchar)y; } bgr += bgr_step - size.width*3; cmyk += cmyk_step - size.width*4; } } void icvCvt_CMYK2RGB_8u_C4C3R( const uchar* cmyk, int cmyk_step, uchar* rgb, int rgb_step, Size size ) { int i; for( ; size.height--; ) { for( i = 0; i < size.width; i++, rgb += 3, cmyk += 4 ) { int c = cmyk[0], m = cmyk[1], y = cmyk[2], k = cmyk[3]; c = k - ((255 - c)*k>>8); m = k - ((255 - m)*k>>8); y = k - ((255 - y)*k>>8); rgb[0] = (uchar)c; rgb[1] = (uchar)m; rgb[2] = (uchar)y; } rgb += rgb_step - size.width*3; cmyk += cmyk_step - size.width*4; } } void icvCvt_CMYK2Gray_8u_C4C1R( const uchar* cmyk, int cmyk_step, uchar* gray, int gray_step, Size size ) { int i; for( ; size.height--; ) { for( i = 0; i < size.width; i++, cmyk += 4 ) { int c = cmyk[0], m = cmyk[1], y = cmyk[2], k = cmyk[3]; c = k - ((255 - c)*k>>8); m = k - ((255 - m)*k>>8); y = k - ((255 - y)*k>>8); int t = descale( y*cB + m*cG + c*cR, SCALE ); gray[i] = (uchar)t; } gray += gray_step; cmyk += cmyk_step - size.width*4; } } void CvtPaletteToGray( const PaletteEntry* palette, uchar* grayPalette, int entries ) { int i; for( i = 0; i < entries; i++ ) { icvCvt_BGR2Gray_8u_C3C1R( (uchar*)(palette + i), 0, grayPalette + i, 0, Size(1,1) ); } } void FillGrayPalette( PaletteEntry* palette, int bpp, bool negative ) { int i, length = 1 << bpp; int xor_mask = negative ? 255 : 0; for( i = 0; i < length; i++ ) { int val = (i * 255/(length - 1)) ^ xor_mask; palette[i].b = palette[i].g = palette[i].r = (uchar)val; palette[i].a = 0; } } bool IsColorPalette( PaletteEntry* palette, int bpp ) { int i, length = 1 << bpp; for( i = 0; i < length; i++ ) { if( palette[i].b != palette[i].g || palette[i].b != palette[i].r ) return true; } return false; } uchar* FillUniColor( uchar* data, uchar*& line_end, int step, int width3, int& y, int height, int count3, PaletteEntry clr ) { do { uchar* end = data + count3; if( end > line_end ) end = line_end; count3 -= (int)(end - data); for( ; data < end; data += 3 ) { WRITE_PIX( data, clr ); } if( data >= line_end ) { line_end += step; data = line_end - width3; if( ++y >= height ) break; } } while( count3 > 0 ); return data; } uchar* FillUniGray( uchar* data, uchar*& line_end, int step, int width, int& y, int height, int count, uchar clr ) { do { uchar* end = data + count; if( end > line_end ) end = line_end; count -= (int)(end - data); for( ; data < end; data++ ) { *data = clr; } if( data >= line_end ) { line_end += step; data = line_end - width; if( ++y >= height ) break; } } while( count > 0 ); return data; } uchar* FillColorRow8( uchar* data, uchar* indices, int len, PaletteEntry* palette ) { uchar* end = data + len*3; while( (data += 3) < end ) { *((PaletteEntry*)(data-3)) = palette[*indices++]; } PaletteEntry clr = palette[indices[0]]; WRITE_PIX( data - 3, clr ); return data; } uchar* FillGrayRow8( uchar* data, uchar* indices, int len, uchar* palette ) { int i; for( i = 0; i < len; i++ ) { data[i] = palette[indices[i]]; } return data + len; } uchar* FillColorRow4( uchar* data, uchar* indices, int len, PaletteEntry* palette ) { uchar* end = data + len*3; while( (data += 6) < end ) { int idx = *indices++; *((PaletteEntry*)(data-6)) = palette[idx >> 4]; *((PaletteEntry*)(data-3)) = palette[idx & 15]; } int idx = indices[0]; PaletteEntry clr = palette[idx >> 4]; WRITE_PIX( data - 6, clr ); if( data == end ) { clr = palette[idx & 15]; WRITE_PIX( data - 3, clr ); } return end; } uchar* FillGrayRow4( uchar* data, uchar* indices, int len, uchar* palette ) { uchar* end = data + len; while( (data += 2) < end ) { int idx = *indices++; data[-2] = palette[idx >> 4]; data[-1] = palette[idx & 15]; } int idx = indices[0]; uchar clr = palette[idx >> 4]; data[-2] = clr; if( data == end ) { clr = palette[idx & 15]; data[-1] = clr; } return end; } uchar* FillColorRow1( uchar* data, uchar* indices, int len, PaletteEntry* palette ) { uchar* end = data + len*3; const PaletteEntry p0 = palette[0], p1 = palette[1]; while( (data += 24) < end ) { int idx = *indices++; *((PaletteEntry*)(data - 24)) = (idx & 128) ? p1 : p0; *((PaletteEntry*)(data - 21)) = (idx & 64) ? p1 : p0; *((PaletteEntry*)(data - 18)) = (idx & 32) ? p1 : p0; *((PaletteEntry*)(data - 15)) = (idx & 16) ? p1 : p0; *((PaletteEntry*)(data - 12)) = (idx & 8) ? p1 : p0; *((PaletteEntry*)(data - 9)) = (idx & 4) ? p1 : p0; *((PaletteEntry*)(data - 6)) = (idx & 2) ? p1 : p0; *((PaletteEntry*)(data - 3)) = (idx & 1) ? p1 : p0; } int idx = indices[0]; for( data -= 24; data < end; data += 3, idx += idx ) { const PaletteEntry clr = (idx & 128) ? p1 : p0; WRITE_PIX( data, clr ); } return data; } uchar* FillGrayRow1( uchar* data, uchar* indices, int len, uchar* palette ) { uchar* end = data + len; const uchar p0 = palette[0], p1 = palette[1]; while( (data += 8) < end ) { int idx = *indices++; *((uchar*)(data - 8)) = (idx & 128) ? p1 : p0; *((uchar*)(data - 7)) = (idx & 64) ? p1 : p0; *((uchar*)(data - 6)) = (idx & 32) ? p1 : p0; *((uchar*)(data - 5)) = (idx & 16) ? p1 : p0; *((uchar*)(data - 4)) = (idx & 8) ? p1 : p0; *((uchar*)(data - 3)) = (idx & 4) ? p1 : p0; *((uchar*)(data - 2)) = (idx & 2) ? p1 : p0; *((uchar*)(data - 1)) = (idx & 1) ? p1 : p0; } int idx = indices[0]; for( data -= 8; data < end; data++, idx += idx ) { data[0] = (idx & 128) ? p1 : p0; } return data; } } // namespace