Manhattan Distance based Nearest Vicinity Guided Filter Design for High Density Salt and Pepper Noise Removal

Abstract

Elimination of highly densified salt and pepper noise from digital images is an exigent business in image processing domain. Numerous state-of-art filters have shown judicious outcomes at low and average noise concentrations. In this paper a nearest vicinity guided two-step impulse noise filter (NVGINF) is proposed. The detection phase segregates the un-corrupted pixels by neglecting the ‘0’ (pepper noise) and ‘255’ (salt noise) pixel intensities from the images. In the removal phase, the corrupted pixels are restored by substituting them with detected un-corrupted pixels located at least Manhattan distance, using equivalent mean-median measures. NVGINF is assessed upon three test images using Peak Signal to Noise Ratio (PSNR), Structural Similarity Index Measurement (SSIM) and Average Run Time (ART). It has been witnessed that the proposed NVGINF demonstrates viable outcome up against a number of contemporary filters specifically at high noise densities.