Design of a hierarchical fuzzy filter for removal of heavy impulse noise from corrupted images

In this paper we discuss the design of a digital image enhancement system based on a hierarchical fuzzy logic (HFL) approach. Hierarchical fussy systems, first introduced in [G.V.S Raju and J. Zhou (1993)] are capable of substantially reducing the number of fuzzy rules to be learnt. We show how evolutionary algorithms (EAs) can be used to learn the fuzzy rules in a fuzzy image filter as opposed to determining the rules using human intuition. Results are presented for the well-known ´Lena´ image and another ´hill´ image to prove that the newly designed hierarchical filter had acquired sufficient knowledge to enhance images which were not used during the training phase of the algorithm.