A noise-exclusive adaptive filtering framework for removing impulsenoise in digital images

摘要

A class of noise-exclusive adaptive filters for removing impulsennoise from digital images is developed and analyzed in this brief. Thenfiltering scheme is based on noise detection using a self-organizingnneural network and noise excluding estimation. These filters suppressnimpulse noise effectively while preserving fine image details.nApplications of the filters to several images show that their propertiesnof efficient impulse noise suppression, edges and fine detailnpreservation, minimum signal distortion, or minimum mean square errornare better than those of the traditional median-type filters