Cellular automata-based approach for salt-and-pepper noise filtration

Application of cellular automata (CA) to digital image processing has achieved considerable attention in last several years. CA are now employed in noise filtration of digital images; particularly many CA-based impulse noise filters have been proposed. Salt-and-pepper noise is a special kind of impulse noise and is introduced in an image during transmission over transmission media by external noise sources like atmospheric disturbances, or due to corrupted hardware memory locations or fault in camera sensors. In this paper, we present five salt-and-pepper noise filters based on modifications of outer totalistic cellular automata (OTCA) with the adaptive neighborhood. Use of OTCA model makes the proposed filters computationally simple on one hand and the use of adaptive neighborhood help the filters to provide efficient noise filtration at varying noise densities on the other. Comparative analysis of these filters followed by comparison with several standard and CA-based filters in terms of peak signal to noise ratio (PSNR) and structural similarity (SSIM) index is presented. (c) 2018 The Authors. Production and hosting by Elsevier B.V. on behalf of King Saud University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).