A fuzzy rule based adaptive centre weighted median filter

In this paper, a fuzzy rule based adaptive centre weighted median (FCWM) filter for the image restoration is introduced. According to some statistic measures of the input image, the parameters of fuzzy membership functions are determined to construct a set of fuzzy rules. The adequate centre weight is the output of fuzzy rules in order to attain the final output pixel, which is adaptive with the sliding of the local filtering window in raster scanning order across the image. Unlike the previous detection based filters looking for the threshold iteratively, the proposed filter establishes the associated rules quickly and in a straightforward manner. A lot of computer simulations show that the proposed FCWM filter can achieve a superior performance either in detail preservation or in noise suppression.