An improved adaptive impulsive noise suppression scheme for digital images

This paper proposes an improved method for suppressing impulsive noise of random values from contaminated images. Filtering is carried out only on noisy pixels keeping the noise-free ones intact. To decide whether a pixel is corrupted or not, the proposed scheme employs an artificial neural network with parameters of the noisy image as input. Filtering is performed recursively so that the restored noisy pixels of the current window participate in the detection phase of the next window. Exhaustive simulations indicate that the proposed scheme consistently outperforms its counterparts in suppressing impulsive noise and retaining original image properties.