OpenCL
output=imdilate_ATI(input);
Both output and input are float types. Currently input values must be between 0 and 255. strel('disk',4,0) is only supported currently.
Dilation take the maximum of each input pixel's neighbors as the corresponding output pixel. Since nearby input pixels
may have common pixels, among which only one maximum exists.Besides using vector type and shared memory for better bandwidth
utilization, one key optimization technique we use is to re-use the overlapped computation among several nearby input pixels. This way
many comparison computations can be saved.