A noise smoother using cascaded FIRE filters
F. Russo, Giovanni Ramponi · 2002
FIRE (Fuzzy Inference Ruled by Else-action) operators are a family of rule-based fuzzy operators dedicated to image processing applications. In this work the latest member of this family is presented: a noise smoother which is composed of two cascaded FIRE filters. Such operators are based on an improved inference mechanism which permits a better control of the fuzzy set parameters in order to achieve the desired performance. As a result, cascaded operators can be very easily tuned in order to obtain a more efficient design of a fuzzy smoother. The performance of the proposed method is analyzed for the case of images degraded by different levels of noise and the obtained results compare favourably to those yielded by other methods in the literature.>