Mid Point Weighted Median Filter for Noise Removal and Smoothening of Images

Abstract

In this paper, a Weighted Median Filter (WMF) mainly Mid-Point WMF (MP WMF) is used to smoothen the images or to remove the impulsive noise present in the input images. It is an improvised version of the median filter. This weighted filter uses the basic concept of the median filter with the enhancement in their mask from normalized mask to the weighted mask. In this weighted median filter, the weight-age is giving to the mid-point pixel i.e. centre pixels or its neighbouring pixels by increasing the repetition of those pixels so that the weight of those mid-point pixel and neighbour pixels get increased. It is shown that the weighted median filter performs more accurately than the normal median filter. Some relationship between MP WMF and other two Four Neighbouring (N4) and Diagonal Neighbouring (ND) Weighted Median Filter are derived. Experimental work has demonstrated that the MP WMF outperform very well than the standard median filters in terms of their noise reduction and in smoothening the image. The proposed MP WMF also provides very good robustness for impulsive noise.