Decision-based neighborhood-referred unsymmetrical trimmed variants filter for the removal of high-density salt-and-pepper noise in images and videos

Abstract

A fixed $$3\times 3$$ decision-based algorithm is proposed for the enhancement of images, and videos that are heavily corrupted by salt-and-pepper noise are proposed. The algorithm uses unsymmetrical trimmed variants for the noise removal. The corrupted pixel is replaced based on the number of non-noisy pixel in the current processing window. The proposed algorithm was applied on various grayscale, and videos that gave excellent peak signal-to-noise ratio, high image enhancement factor, low mean square error, and very good SSIM with excellent edge preservation even at high noise densities. If all the pixels of the current processing window are noisy, then instead of unsymmetrical midpoint, global trimmed mean of the image is replaced as output. The proposed algorithm shows excellent noise suppression capability, when compared to standard and existing filters in terms of both qualitative and quantitative measures at highly noisy environment.