Images enhancement using hybrid Median-Mean filters

Abstract

This paper presents new techniques for noise reduction depending on filters. The two proposed filters are to simulate and improve the effects of two kinds of smoothing filters: Mean and Median filters. These two filters are of amplitude and statistic kinds but they are vibrated between the linear and nonlinear kinds.The first proposed filter takes the average of mean and median filters for each pixel in the image as a new pixel in the filtered image. While the second filter takes the average of the W median intensity values in the mask (H×W) for each pixel in the image as a new pixel in the filtered image.Simulations show that the proposed filters work well in suppressing two different kinds of noise: Poisson and Gaussian from both images grayscale and RGB while preserving image details.