Impulse noise detection and filtering using fuzzy models

A new technique for signal denoising is presented The method is specifically designed to perform a very accurate filtering of signals corrupted by random-valued impulse noise. The proposed approach adopts fuzzy models in order to estimate the degree of waveform corruption and then apply the appropriate noise correction. The parallel filtering architecture combines effective noise removal and preservation of the signal waveform. Computer simulations show that the proposed method gives better results than using conventional techniques.