Fast impulse noise suppression in colour images using most median filter

This paper introduces a noise detection akgorithm, Most Median Filter (MMF) for the detection and removal of impulse noise from colour images. The proposed filter has die potential for increased detail preservation as compared to the existing benchmark filters e.g. Vector Median Filter (VMF), Vector Directional Filter (VDF) and Directional Distance Filter (DDF). In the operation of the proposed algorithm, each component of an RGB image is processed for noise detection. After a noise pixel is detected , this algorithm then uses VMF, VDF or DDF, based on the heuristics, for noise removal. The effectiveness of this algorithm is evaluated using two well known matrices namely Peak Signal to Noise Ratio (PSNR) and Normalized Colour Difference (NCD). Along with these evaluating criteria, we have also measured the performance of these algorithms in terms of their Computation complexities.