Local Criteria : A Unified Approach to Locally Adaptive Linear and Rank Filters

Abstract

During the last several years a lot of new linear and nonlinear rank algorithms for signal and picture processing have been proposed [1-13]. The linear ones have been derived from one or another statistical model of the signal to be processed. The rank ones have been substantiated from different positions and without any rather general notion of optimality. The aim of this paper is to introduce a new family of criteria of processing quality as a basis for unified approach to synthesis of optimal linear and rank filters for signal and picture processing.