Adaptive mean and median filters for image restoration
B. Yegnanarayana, PSrinivasa Rao, K.J. Palaniswamy · 1st IASTED International Symposium on Signal Processing and its Applications · 1987
We report here our studies on new adaptive mean and median filters for filtering noise from images. These filters replace the central pixel of a sliding window by the mean (median) of the effective pixels in the window. The effective pixels are identified by making use of a threshold value, which is the defining parameter of the adaptive filters. The dependence of the performance of these filters on the value of the threshold is discussed. It is shown that these adaptive filters preserve the edges and impulses in images while filtering out the noise. Performance of the adaptive filters is better than conventional filters for low noise levels and for large window sizes. In the worst case of high noise levels, the performance of the adaptive filters approaches that of the conventional filters. The results are illustrated with several processed noisy images.