Adaptive Filtering of Images with Filter's Aggregation in the Framework of Evidence Theory

One of the most important problems of image processing is the task of pre-cleaning them from noise. There are many well-accepted methods for image filtering. However, along with their advantages, they have their drawbacks. Thus, the task of combining several filters into one filter seems to be relevant. The problem of constructing an aggregating filter with the use of tools of evidence theory (the Dempster-Shafer theory) is considered in this paper. The efficiency of constructing such an operator with using various rules for combining evidences and considering their discounting is investigated. Experimental testing was conducted for various types of noise. The comparative analysis of the efficiency of image filtering with aggregation filters with the classical filtering methods was carried out with respect to the various cost functionals. Image Processing; Median Filtering; Adaptive Filtering; Evidence theory.