Parallelization of Fuzzy-Classical Filters For Image Noise Reduction

Several fuzzy filters for image noise reduction have already been developed. In general, they are able to preserve images in a more comprehensive means than classical filters, and they have the ability to combine edge-preservation and smoothing. However, the implementation of fuzzy filters is very time- consuming. On the other hand, parallel and grid computing technologies are efficient tools for implementing fuzzy filters. In this paper, we propose a parallel skeleton for some fuzzy weighted mean filters. We have implemented the algorithms using MatlabMPI (a parallel, message passing version of Matlab). Our experiments show the feasibility and efficiency of the algorithms.