The use of sample selection probabilities for stack filter design

Abstract

We propose a procedure for stack filter design that takes into consideration the filter's sample selection probabilities. A statistical optimization of stack filters can result in a class of stack filters, all of which are statistically equivalent. Such a situation arises in cases of nonsymmetric noise distributions or in the presence of constraints. Among the set of equivalent stack filters, our method constructs a statistically optimal stack filter whose sample selection probabilities are concentrated in the center of its window. This leads to improvement of detail preservation.