Vector sigma filters for noise detection and removal in color images

This paper presents a new adaptive filtering approach capable of detecting and removing impulsive noise in multichannel images. The proposed methodology constitutes a powerful unified framework for multichannel signal processing. Robust order-statistic concepts and statistical measure of vectors' deviation are used in conjunction with different distance measures among multichannel inputs to determine an efficient switching rule between filter output and no filtering (identity operation). The special case of color image filtering is studied as an important example of multichannel signal processing. Simulation studies reported in this paper indicate that the proposed filter class is computationally attractive, has excellent performance, and is able to preserve fine details while suppressing impulsive noise. (c) 2005 Elsevier Inc. All rights reserved.