Fuzzy bilateral filtering for color images

Bilateral filtering is a well-known technique for smoothing gray-scale and color images while preserving edges and image details by means of an appropriate nonlinear combination of the color vectors in a neighborhood. The pixel colors are combined based on their spatial closeness and photometric similarity. In this paper, a particular class of fuzzy metrics is used to represent the spatial and photometric relations between the color pixels adapting the classical bilateral filtering. It is shown that the use of these fuzzy metrics is more appropriate than the classical measures used.