Advanced impulse detection based on pixel-wise MAD

Abstract

A new method for impulse noise removal is presented, where a robust estimator of the variance, MAD (median of the absolute deviations from the median), is modified and used to efficiently separate noisy pixels from the image details. The algorithm is free of varying parameters, requires no previous training or optimization, and successfully removes all types of impulse noise. The pixel-wise MAD concept is straightforward, low in complexity, and achieves high filtering performance. © 2004 IEEE.