Impulsive noise detection by double noise detector and removal using adaptive neural-fuzzy inference system

A novel impulsive noise elimination filter combining a double noise detector with an adaptive neural-fuzzy inference system (ANFIS) is proposed. The proposed filter has a two-phase scheme for removing impulse noise: detecting noisy pixels and removing noise. In the detecting noisy pixels phase, noise candidates identified with the noise detection algorithm of the adaptive median filter are judged again by local fuzzy membership function to improve the accurate rate of noise detection. In the removing noise phase, the corrupted pixels are restored using adaptive neural-fuzzy inference system based on their neighbor uncorrupted pixels. Experimental results show that the proposed filter yields a better performance than the recent filters in terms of details preservation and noise suppression, and our scheme can remove impulsive noise with noise level as high as 90%.