A fast algorithm for point processing

Point processing is one of the most simple and effective methods in image enhancement. Usually, point processing transforms an input image f(i,j) into g(i,j) through a mapping T(*) point by point, that is, g(i,j)=T(f(i,j)). Here, (i,j) is a pixel in these images. This paper puts forward one way of a fast template lookup table algorithm for point processing whose major computations are focused on 256 templates. Compared with the point by point algorithm, the fast template lookup table algorithm may theoretically reach a speedup ratio of about N² /256 for a N×N pixel image