Image Segmentation
Suppose it is known a priori that the image consists of only two principal brightness levels with prior probabilities P1 and P2.
Consider the situation where natural variations or noise modify the two gray levels to distributions represented by Gaussian PDFs p1(x) and p2(x), where x represents the gray level.
The PDF of the image gray levels is then
where, ?1 and ?2 are the means of the two brightness regions, and ? 1 and ? 2 are their standard deviations.