Experimental results of an evolution-based adaptation strategy for VQ image filtering

In this paper, we propose the filtering of images based on the codebooks ob tained from an evolution-based adaptation strategy for vector quantization (VQ). This evolution-based VQ Bayesian filter (VQBF) is applied to noise re moval and segmentation of a high-resolution magnetic resonance image. We co mpare our approach with other more conventional smoothing filters. The resu lts show that VQBF performs a smoothing that preserves region boundaries an d small details. It does not show the strong boundary diffusion and displac ement that are common to smoothing filters. Border detection on the filtere d images is also presented. (C) 2001 Elsevier Science Inc. All rights reser ved.