|
Java example source code file (mlib_v_ImageChannelInsert.c)
The mlib_v_ImageChannelInsert.c Java example source code/* * Copyright (c) 1998, 2003, Oracle and/or its affiliates. All rights reserved. * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * * This code is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License version 2 only, as * published by the Free Software Foundation. Oracle designates this * particular file as subject to the "Classpath" exception as provided * by Oracle in the LICENSE file that accompanied this code. * * This code is distributed in the hope that it will be useful, but WITHOUT * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License * version 2 for more details (a copy is included in the LICENSE file that * accompanied this code). * * You should have received a copy of the GNU General Public License version * 2 along with this work; if not, write to the Free Software Foundation, * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. * * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA * or visit www.oracle.com if you need additional information or have any * questions. */ /* * FUNCTIONS * mlib_ImageChannelInsert - Copy the source image into the selected * channels of the destination image * * SYNOPSIS * mlib_status mlib_ImageChannelInsert(mlib_image *dst, * mlib_image *src, * mlib_s32 cmask); * * ARGUMENT * dst Pointer to destination image. * src Pointer to source image. * cmask Destination channel selection mask. * The least significant bit (LSB) is corresponding to the * last channel in the destination image data. * The bits with value 1 stand for the channels selected. * If more than N channels are selected, the leftmost N * channels are inserted, where N is the number of channels * in the source image. * * RESTRICTION * The src and dst must have the same width, height and data type. * The src and dst can have 1, 2, 3 or 4 channels. * The src and dst can be either MLIB_BYTE, MLIB_SHORT, MLIB_INT, * MLIB_FLOAT or MLIB_DOUBLE. * * DESCRIPTION * Copy the source image into the selected channels of the destination * image */ #include <stdlib.h> #include "mlib_image.h" #include "mlib_ImageCheck.h" /***************************************************************/ /* functions defined in mlib_v_ImageChannelInsert_1.c */ void mlib_v_ImageChannelInsert_U8(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 channels, mlib_s32 channeld, mlib_s32 width, mlib_s32 height, mlib_s32 cmask); void mlib_v_ImageChannelInsert_D64(mlib_d64 *src, mlib_s32 slb, mlib_d64 *dst, mlib_s32 dlb, mlib_s32 channels, mlib_s32 channeld, mlib_s32 width, mlib_s32 height, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 channels, mlib_s32 channeld, mlib_s32 width, mlib_s32 height, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S32(mlib_s32 *src, mlib_s32 slb, mlib_s32 *dst, mlib_s32 dlb, mlib_s32 channels, mlib_s32 channeld, mlib_s32 width, mlib_s32 height, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_12_A8D1X8(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_12_A8D2X8(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_12_D1(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_12(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_13_A8D1X8(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_13_A8D2X8(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_13_D1(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_13(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_14_A8D1X8(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_14_A8D2X8(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_14_D1(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_U8_14(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_12_A8D1X4(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_12_A8D2X4(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_12_D1(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_12(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_13_A8D1X4(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_13_A8D2X4(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_13_D1(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_13(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_14_A8D1X4(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_14_A8D2X4(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_14_D1(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize, mlib_s32 cmask); void mlib_v_ImageChannelInsert_S16_14(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize, mlib_s32 cmask); /***************************************************************/ /* functions defined in mlib_v_ImageChannelInsert_34.c */ void mlib_v_ImageChannelInsert_U8_34R_A8D1X8(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize); void mlib_v_ImageChannelInsert_U8_34R_A8D2X8(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize); void mlib_v_ImageChannelInsert_U8_34R_D1(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize); void mlib_v_ImageChannelInsert_U8_34R(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize); void mlib_v_ImageChannelInsert_S16_34R_A8D1X4(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize); void mlib_v_ImageChannelInsert_S16_34R_A8D2X4(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize); void mlib_v_ImageChannelInsert_S16_34R_D1(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize); void mlib_v_ImageChannelInsert_S16_34R(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize); void mlib_v_ImageChannelInsert_U8_34L_A8D1X8(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize); void mlib_v_ImageChannelInsert_U8_34L_A8D2X8(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize); void mlib_v_ImageChannelInsert_U8_34L_D1(mlib_u8 *src, mlib_u8 *dst, mlib_s32 dsize); void mlib_v_ImageChannelInsert_U8_34L(mlib_u8 *src, mlib_s32 slb, mlib_u8 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize); void mlib_v_ImageChannelInsert_S16_34L_A8D1X4(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize); void mlib_v_ImageChannelInsert_S16_34L_A8D2X4(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize); void mlib_v_ImageChannelInsert_S16_34L_D1(mlib_s16 *src, mlib_s16 *dst, mlib_s32 dsize); void mlib_v_ImageChannelInsert_S16_34L(mlib_s16 *src, mlib_s32 slb, mlib_s16 *dst, mlib_s32 dlb, mlib_s32 xsize, mlib_s32 ysize); /***************************************************************/ #ifdef MLIB_TEST mlib_status mlib_v_ImageChannelInsert(mlib_image *dst, mlib_image *src, mlib_s32 cmask) #else mlib_status mlib_ImageChannelInsert(mlib_image *dst, mlib_image *src, mlib_s32 cmask) #endif { const mlib_s32 X8 = 0x7; const mlib_s32 X4 = 0x3; const mlib_s32 X2 = 0x1; const mlib_s32 A8D1 = MLIB_IMAGE_ALIGNED8 | MLIB_IMAGE_ONEDVECTOR; const mlib_s32 A8D2X8 = MLIB_IMAGE_ALIGNED8 | MLIB_IMAGE_STRIDE8X | MLIB_IMAGE_WIDTH8X; const mlib_s32 A8D2X4 = MLIB_IMAGE_ALIGNED8 | MLIB_IMAGE_STRIDE8X | MLIB_IMAGE_WIDTH4X; const mlib_s32 A8D2X2 = MLIB_IMAGE_ALIGNED8 | MLIB_IMAGE_STRIDE8X | MLIB_IMAGE_WIDTH2X; void *sp; /* pointer for pixel in src */ void *dp; /* pointer for pixel in dst */ mlib_s32 ncmask = 0; /* normalized channel mask */ mlib_s32 channels; /* number of channels for src */ mlib_s32 channeld; /* number of channels for dst */ mlib_s32 width, height;/* for src and dst */ mlib_s32 strides; /* strides in bytes for src */ mlib_s32 strided; /* strides in bytes for dst */ mlib_s32 flags; mlib_s32 flagd; mlib_s32 dsize; int i, bit1count = 0; MLIB_IMAGE_CHECK(src); MLIB_IMAGE_CHECK(dst); MLIB_IMAGE_TYPE_EQUAL(src,dst); MLIB_IMAGE_SIZE_EQUAL(src,dst); channels = mlib_ImageGetChannels(src); channeld = mlib_ImageGetChannels(dst); width = mlib_ImageGetWidth(src); height = mlib_ImageGetHeight(src); strides = mlib_ImageGetStride(src); strided = mlib_ImageGetStride(dst); sp = mlib_ImageGetData(src); dp = mlib_ImageGetData(dst); flags = mlib_ImageGetFlags(src); flagd = mlib_ImageGetFlags(dst); dsize = width * height; /* normalize the cmask, and count the number of bit with value 1 */ for (i = (channeld - 1); i >= 0; i--) { if (((cmask & (1 << i)) != 0) && (bit1count < channels)) { ncmask += (1 << i); bit1count++; } } /* do not support the cases in which the number of selected channels is * less than the nubmber of channels in the source image */ if (bit1count < channels) { return MLIB_FAILURE; } if (((channels == 1) && (channeld == 1)) || ((channels == 2) && (channeld == 2)) || ((channels == 3) && (channeld == 3)) || ((channels == 4) && (channeld == 4))) { return mlib_ImageCopy(dst, src); } switch (mlib_ImageGetType(src)) { case MLIB_BYTE: if (channels == 1) { switch (channeld) { case 2: if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X8) == 0)) { mlib_v_ImageChannelInsert_U8_12_A8D1X8((mlib_u8 *)sp, (mlib_u8 *)dp, dsize, ncmask); } else if (((flags & A8D2X8) == 0) && ((flagd & A8D2X8) == 0)) { mlib_v_ImageChannelInsert_U8_12_A8D2X8((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height, ncmask); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_U8_12_D1((mlib_u8 *)sp, (mlib_u8 *)dp, dsize, ncmask); } else { mlib_v_ImageChannelInsert_U8_12((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height, ncmask); } break; case 3: if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X8) == 0)) { mlib_v_ImageChannelInsert_U8_13_A8D1X8((mlib_u8 *)sp, (mlib_u8 *)dp, dsize, ncmask); } else if (((flags & A8D2X8) == 0) && ((flagd & A8D2X8) == 0)) { mlib_v_ImageChannelInsert_U8_13_A8D2X8((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height, ncmask); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_U8_13_D1((mlib_u8 *)sp, (mlib_u8 *)dp, dsize, ncmask); } else { mlib_v_ImageChannelInsert_U8_13((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height, ncmask); } break; case 4: if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X8) == 0)) { mlib_v_ImageChannelInsert_U8_14_A8D1X8((mlib_u8 *)sp, (mlib_u8 *)dp, dsize, ncmask); } else if (((flags & A8D2X8) == 0) && ((flagd & A8D2X8) == 0)) { mlib_v_ImageChannelInsert_U8_14_A8D2X8((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height, ncmask); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_U8_14_D1((mlib_u8 *)sp, (mlib_u8 *)dp, dsize, ncmask); } else { mlib_v_ImageChannelInsert_U8_14((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height, ncmask); } break; default: return MLIB_FAILURE; } } else { if ((channels == 3) && (channeld == 4) && (ncmask == 7)) { if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X8) == 0)) { mlib_v_ImageChannelInsert_U8_34R_A8D1X8((mlib_u8 *)sp, (mlib_u8 *)dp, dsize); } else if (((flags & A8D2X8) == 0) && ((flagd & A8D2X8) == 0)) { mlib_v_ImageChannelInsert_U8_34R_A8D2X8((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_U8_34R_D1((mlib_u8 *)sp, (mlib_u8 *)dp, dsize); } else { mlib_v_ImageChannelInsert_U8_34R((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height); } } else if ((channels == 3) && (channeld == 4) && (ncmask == 14)) { if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X8) == 0)) { mlib_v_ImageChannelInsert_U8_34L_A8D1X8((mlib_u8 *)sp, (mlib_u8 *)dp, dsize); } else if (((flags & A8D2X8) == 0) && ((flagd & A8D2X8) == 0)) { mlib_v_ImageChannelInsert_U8_34L_A8D2X8((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_U8_34L_D1((mlib_u8 *)sp, (mlib_u8 *)dp, dsize); } else mlib_v_ImageChannelInsert_U8_34L((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, width, height); } else { mlib_v_ImageChannelInsert_U8((mlib_u8 *)sp, strides, (mlib_u8 *)dp, strided, channels, channeld, width, height, ncmask); } } break; case MLIB_SHORT: if (channels == 1) { switch (channeld) { case 2: if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X4) == 0)) { mlib_v_ImageChannelInsert_S16_12_A8D1X4((mlib_s16 *)sp, (mlib_s16 *)dp, dsize, ncmask); } else if (((flags & A8D2X4) == 0) && ((flagd & A8D2X4) == 0)) { mlib_v_ImageChannelInsert_S16_12_A8D2X4((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height, ncmask); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_S16_12_D1((mlib_s16 *)sp, (mlib_s16 *)dp, dsize, ncmask); } else { mlib_v_ImageChannelInsert_S16_12((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height, ncmask); } break; case 3: if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X4) == 0)) { mlib_v_ImageChannelInsert_S16_13_A8D1X4((mlib_s16 *)sp, (mlib_s16 *)dp, dsize, ncmask); } else if (((flags & A8D2X4) == 0) && ((flagd & A8D2X4) == 0)) { mlib_v_ImageChannelInsert_S16_13_A8D2X4((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height, ncmask); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_S16_13_D1((mlib_s16 *)sp, (mlib_s16 *)dp, dsize, ncmask); } else { mlib_v_ImageChannelInsert_S16_13((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height, ncmask); } break; case 4: if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X4) == 0)) { mlib_v_ImageChannelInsert_S16_14_A8D1X4((mlib_s16 *)sp, (mlib_s16 *)dp, dsize, ncmask); } else if (((flags & A8D2X4) == 0) && ((flagd & A8D2X4) == 0)) { mlib_v_ImageChannelInsert_S16_14_A8D2X4((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height, ncmask); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_S16_14_D1((mlib_s16 *)sp, (mlib_s16 *)dp, dsize, ncmask); } else { mlib_v_ImageChannelInsert_S16_14((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height, ncmask); } break; default: return MLIB_FAILURE; } } else if ((channels == 3) && (channeld == 4) && (ncmask == 7)) { if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X4) == 0)) { mlib_v_ImageChannelInsert_S16_34R_A8D1X4((mlib_s16 *)sp, (mlib_s16 *)dp, dsize); } else if (((flags & A8D2X4) == 0) && ((flagd & A8D2X4) == 0)) { mlib_v_ImageChannelInsert_S16_34R_A8D2X4((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_S16_34R_D1((mlib_s16 *)sp, (mlib_s16 *)dp, dsize); } else { mlib_v_ImageChannelInsert_S16_34R((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height); } } else if ((channels == 3) && (channeld == 4) && (ncmask == 14)) { if (((flags & A8D1) == 0) && ((flagd & A8D1) == 0) && ((dsize & X4) == 0)) { mlib_v_ImageChannelInsert_S16_34L_A8D1X4((mlib_s16 *)sp, (mlib_s16 *)dp, dsize); } else if (((flags & A8D2X4) == 0) && ((flagd & A8D2X4) == 0)) { mlib_v_ImageChannelInsert_S16_34L_A8D2X4((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height); } else if (((flags & MLIB_IMAGE_ONEDVECTOR) == 0) && ((flagd & MLIB_IMAGE_ONEDVECTOR) == 0)) { mlib_v_ImageChannelInsert_S16_34L_D1((mlib_s16 *)sp, (mlib_s16 *)dp, dsize); } else { mlib_v_ImageChannelInsert_S16_34L((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, width, height); } } else { mlib_v_ImageChannelInsert_S16((mlib_s16 *)sp, strides, (mlib_s16 *)dp, strided, channels, channeld, width, height, ncmask); } break; case MLIB_INT: mlib_v_ImageChannelInsert_S32((mlib_s32 *)sp, strides, (mlib_s32 *)dp, strided, channels, channeld, width, height, ncmask); break; case MLIB_FLOAT: mlib_v_ImageChannelInsert_S32((mlib_s32 *)sp, strides, (mlib_s32 *)dp, strided, channels, channeld, width, height, ncmask); break; case MLIB_DOUBLE: mlib_v_ImageChannelInsert_D64((mlib_d64 *)sp, strides, (mlib_d64 *)dp, strided, channels, channeld, width, height, ncmask); break; case MLIB_BIT: default: return MLIB_FAILURE; /* MLIB_BIT is not supported here */ } return MLIB_SUCCESS; } /***************************************************************/ Other Java examples (source code examples)Here is a short list of links related to this Java mlib_v_ImageChannelInsert.c source code file: |
... this post is sponsored by my books ... | |
#1 New Release! |
FP Best Seller |
Copyright 1998-2024 Alvin Alexander, alvinalexander.com
All Rights Reserved.
A percentage of advertising revenue from
pages under the /java/jwarehouse
URI on this website is
paid back to open source projects.