mlib_ImageConvMxN - MxN convolution
cc [ flag... ] file... –lmlib [ library... ]
#include <mlib.h>
mlib_status mlib_ImageConvMxN(mlib_image *dst, const mlib_image *src,
const mlib_s32 *kernel, mlib_s32 m, mlib_s32 n, mlib_s32 dm,
mlib_s32 dn, mlib_s32 scale, mlib_s32 cmask, mlib_edge edge);
The mlib_ImageConvMxN() function performs a MxN convolution on the source image by using the user-supplied kernel.
The input image and the output image must have the same image type and have the same number of channels. The unselected channels in the output image are not overwritten. For single-channel images, the channel mask is ignored.
For this convolution, the key element of the convolution kernel is located at (dm, dn) of the kernel matrix.
It uses the following equation:
m-1-dm n-1-dn
dst[x][y][i] = SUM SUM src[x+p][y+q][i]*k[p][q]*2**(-scale)
p=-dm q=-dn
where m ≥ 1, n ≥ 1, 0 ≤ dm < m, 0 ≤ dn < n.
The function takes the following arguments:
Pointer to destination image.
Pointer to source image.
Pointer to the convolution kernel, in row major order.
Width of the convolution kernel. m ≥ 1.
Height of the convolution kernel. n ≥ 1.
X coordinate of the key element in the convolution kernel. 0 ≤ dm < m.
Y coordinate of the key element in the convolution kernel. 0 ≤ dn < n.
Scaling factor.
Channel mask to indicate the channels to be convolved, each bit of which represents a channel in the image. The channels corresponding to bits with a value of 1 are those to be processed. For a single-channel image, the channel mask is ignored.
Type of edge condition. It can be one of the following:
MLIB_EDGE_DST_NO_WRITE MLIB_EDGE_DST_FILL_ZERO MLIB_EDGE_DST_COPY_SRC MLIB_EDGE_SRC_EXTEND
The function returns MLIB_SUCCESS if successful. Otherwise it returns MLIB_FAILURE.
See attributes(5) for descriptions of the following attributes:
|
mlib_ImageConv2x2(3MLIB), mlib_ImageConv2x2_Fp(3MLIB), mlib_ImageConv2x2Index(3MLIB), mlib_ImageConv3x3(3MLIB), mlib_ImageConv3x3_Fp(3MLIB), mlib_ImageConv3x3Index(3MLIB), mlib_ImageConv4x4(3MLIB), mlib_ImageConv4x4_Fp(3MLIB), mlib_ImageConv4x4Index(3MLIB), mlib_ImageConv5x5(3MLIB), mlib_ImageConv5x5_Fp(3MLIB), mlib_ImageConv5x5Index(3MLIB), mlib_ImageConv7x7(3MLIB), mlib_ImageConv7x7_Fp(3MLIB), mlib_ImageConv7x7Index(3MLIB), mlib_ImageConvKernelConvert(3MLIB), mlib_ImageConvMxN_Fp(3MLIB), mlib_ImageConvMxNIndex(3MLIB), mlib_ImageConvolveMxN(3MLIB), mlib_ImageConvolveMxN_Fp(3MLIB), mlib_ImageSConv3x3(3MLIB), mlib_ImageSConv3x3_Fp(3MLIB), mlib_ImageSConv5x5(3MLIB), mlib_ImageSConv5x5_Fp(3MLIB), mlib_ImageSConv7x7(3MLIB), mlib_ImageSConv7x7_Fp(3MLIB), mlib_ImageSConvKernelConvert(3MLIB), attributes(5)