A Novel Algorithm for Removing Impulse Noise From Highly Corrupted Image

Abstract

Various switching-based median filter have been proposed in the literature for restoration of extremely
corrupted images by impulsive noise (Salt and pepper noise). Progressive switching median filter (PSMF)
and Boundary discriminative noise detection (BDND) median filter are some of these. It is well known that
standard median filter(SMF) is not suitable for removing impulse noise from highly corrupted images. The
poor efficiency of this filter is due to the fact that it is not able to discriminate the noisy pixel from noiseless
pixel. The proposed method uses noise detection stage and then adaptive window shaped filtering stage for
the restoration of images contaminated by impulse noise. The performance of the proposed method is
demonstrated through computer simulation in comparison with the PSMF method and standard median
filter.