A novel fast algorithm for image median filtering based on improved sign test method

Abstract

This paper studies the median filtering algorithm of digital image and several other related fast algorithms for it, as well as sign test method, then gives an improved algorithm for sign test method, and a new fast median filtering algorithm based on improved sign test method. The new median filtering algorithm based on improved sign test method does not need to sort any of the pixel values in filtering neighbourhood so as to greatly speed up the process of median filtering. At the same time, this algorithm introduced a relative value as the statistic to solve the shortcomings of original sign test method, especially when samples of the two collectivities are almost the same. Algorithm analysis and a lot of experiment results show that the new algorithm can not only speed up the process of image median filtering, but also can maintain much more information of image, especially the edge and texture, compared to several other fast algorithm of image median filtering.