Role of recursive estimation in statistical image enhancement

Abstract

Recursive linear and nonlinear estimation concepts are applied to the problem of image enhancement, whereby procedures for the derivation of the estimators are developed. These procedures are applicable to enhancement of images characterized statistically by their mean and correlation function. Design of an approximate time invariant and a time varying estimator is discussed and the implications of vector scanning are investigated. Feasibility of the approach is illustrated by means of examples, and the computational advantages are presented.