Neural network combined with fuzzy logic to remove salt and pepper noise in digital images

Abstract

Image denoising is an important step in the pre-processing of images. Aim of the paper is to remove the salt and pepper noise on images by using a novel filter based on neural network and fuzzy logic. By this filter it is possible to remove only the pixels that are really affected by noise, thus avoiding image distortion due to the removal of good pixels. A comparison between the proposed filter and the classical median filter shows an increase of about 20% of the peak signal to noise ratio and a better capacity of preserving the details of the images. The proposed approach outperforms the existing algorithms and does not depend on the noise level.