A robust median-mean filter for impulse noise suppression

All the communication channels are prone to noise, which makes it difficult to recover the original data or image at the receiving end. Noise can never be completely suppressed in the channel part, this makes it even more challenging for the engineers to design the receiver that can regenerate the original data stream from the noise affected one. In this paper an impulse noise suppression median-mean filter (INSMF) is designed that works with three different window sizes to reduce the error between original and recovered values. In order to achieve a high degree of accuracy, not only the benefits of median filtering but also Normalized Euclidean Distance and mean value replacement are explored. Theoretical analysis and experimental results are drawn by taking different noise levels present in channel that determines the effectiveness of the proposed filtering method.