Mean-based fast median filter

One of the most popular methods for removing noise in images is to use median filters, with the standard median filter based on the ordering of elements in a set, which is quite slow. This paper presents a fast median filtering algorithm which analyzes neighborhoods to accelerate calculation of the median value of the set by using the mean value. Experimental results show that the algorithm is at least two times faster than the standard median filter and works well with radiological images.