Raul E. Sanchez-Yanez, Horacio Rostro-Gonzalez, and Victor Ayala-Ramirez.
A fuzzy approach for image approximation is discussed. The image is considered as a two input-single output Mamdani-type fuzzy model. The pixel position is the input while the output is the pixel intensity. This fuzzy model is completely obtained from scattered pixel evidence. Such approach can be used for function approximation tasks or as a model identifier from scattered numerical evidence.