Decision based unsymmetrical trimmed mode filter for the removal of salt and pepper noise in images

A Decision based restoration algorithm using Unsymmetrical Trimmed Mode filter for gray scale images that are corrupted by salt and pepper noise is proposed. The processed pixel is examined for 0 or 255; if checked pixel is equal to 0 or 255, then it is considered as noisy pixel else not noisy. If the pixel is noisy and all the four neighbors are noisy then replace the noisy pixel with mean of 4 neighbors. If any of the 4 neighbors is not noisy then replace the corrupted pixel with unsymmetrical trimmed mode. If there are multiple modes then replace the corrupted pixel with smallest of the mode. The uncorrupted pixel is left unchanged. Conventional mode filter does not remove salt and pepper noise even at very low noise densities. The proposed algorithm shows excellent noise suppression capability with good edge preservation in low and medium noisy conditions. The proposed algorithm is compared with standard filters and found to exhibit good quantitative and qualitative results.