A Fuzzy Directional Median Filter for Fixed-value Impulse Noise Removal

Abstract

In this paper, we designed a new fuzzy directional weighted median filter to remove fixed-value impulse noise. Our approach discriminates smooth regions from edges, and apply the median filter on edges along the direction of minimum directional standard deviation. Impulse detection is carried out using a set of fuzzy rules based on directional differences and standard deviation. Objective and subjective evaluations demonstrated that our method is successful in impulse noise removal compared to classical median, and classical directional weighted median filter.