cc [ flag... ] file... -lmlib [ library... ] #include <mlib.h> mlib_status mlib_ImageSubsampleAverage(mlib_image *dst, const mlib_image *src, mlib_d64 xscale, mlib_d64 yscale);
mlib_status mlib_ImageSubsampleAverage_Fp(mlib_image *dst, const mlib_image *src, mlib_d64 xscale, mlib_d64 yscale);
Each function scales an image down with an adaptive box filter.
The subsampling algorithm performs the scaling operation by averaging all the pixel values from a block in the source image that correspond to the destination pixel.
The width and height of the source block for a destination pixel are computed as:
blockX = (int)ceil(1.0/xscale); blockY = (int)ceil(1.0/yscale);
If we denote a pixel's location in an image by its column number and row number (both counted from 0), the destination pixel at (i, j) is backward mapped to the source block whose upper-left corner pixel is at (xValues[i], yValues[j]), where
xValues[i] = (int)(i/xscale + 0.5); yValues[j] = (int)(j/yscale + 0.5);
The width and height of the filled area in the destination are restricted by
dstW = (int)(srcWidth * xscale); dstH = (int)(srcHeight * yscale);
where srcWidth and srcHeight are width and height of the source image.
Since the block size in source is defined from scale factors with roundup, some blocks (the rightmost and the bottommost blocks) may overrun the border of the source image by 1 pixel. In this case, such blocks are moved by 1 pixel to left/up direction in order to be inside of the source image.
The function takes the following arguments:
The function returns MLIB_SUCCESS if successful. Otherwise it returns MLIB_FAILURE.
See attributes(5) for descriptions of the following attributes:
mlib_ImageSubsampleBinaryToGray(3MLIB), mlib_ImageFilteredSubsample(3MLIB), mlib_ImageZoomTranslate(3MLIB), mlib_ImageZoomTranslate_Fp(3MLIB), attributes(5)