Nonlinear filtering of multivariate images under robust error criterion.

Koivunen, V · IEEE Trans Image Process · 1996

basic_science · Level V

Where this comes from

Abstract

A class of nonlinear filters for multivariate data is introduced. A robust error criterion is minimized. Approximate algorithms for computing the filter output are developed. A polynomial signal model is used in applications where the signal amplitude has to be retained with high fidelity. Simulated data and RGB color image data are used in experiments.