FUZZY ADAPTIVE FILTERS FOR MULTICHANNEL IMAGE-PROCESSING

New filter classes for multichannel image processing are introduced an d analyzed. The proposed methodology constitutes a unifying and powerf ul framework for multichannel image processing. The new fillers use fu zzy membership functions based on different distance measures among th e image vectors to adapt to local data in the image. The principle beh ind the new nonlinear filters is explained in detail. Using the propos ed methodology multichannel nonlinear problems are treated from a glob al viewpoint that readily yields and unifies previous, seemingly unrel ated results, The new approach provides insight into the nature of the filtering process and the structure of the underlying nonlinear opera tor. The special case of color image processing is studied as an impor tant example of multichannel image processing. Simulation results indi cate that the new filters are computationally attractive and have exce llent performance.