Image Filtering Using Fuzzy $$S-$$metrics

Abstract

AbstractIn this paper image filtering is performed by fuzzy metrics. The authors concentrated on improving the quality and sharpness of the filtered image, computed by the image quality metrics UIQI and CPBD. Fuzzy metrics and so-called aggregated metrics determine the similarity measure (distance) by the color of the pixels from the window with central pixel, as well as the spatial distance itself. Appropriate selection of the fuzzy complement that defines the \(S-\)metric dual with earlier examined \(T-\)metrics would give the whole spectrum of original metrics with higher sharpness of filtered image. In this paper, combinations of fuzzy \(S-\)metrics are considered, in order to obtain good criteria during the image filtering process, which is explained in more detail in the paper. Higher sharpness index than images filtered with the median filter is obtained in this paper, where the images are filtered with this modified algorithm. The fuzzy metric parameters that produce images with the better quality and sharpness are determined experimentally.KeywordsFuzzy complementFuzzy \(S-\)metricImage filteringImage sharpness