Using High Order Total Variation for Denoising Speckle, Gaussian, Salt & Pepper

Abstract

Noise reduction for image enhancement is the important issue for any image processing algorithm. Total variation (TV) regularization based methods were proposed for multiplicative speckle noise reduction, though sometimes have the undesirable staircase effect. As a solution for this problem, high order TV (High-TV) algorithm was proposed. In this paper we use the High-TV not only for speckle but also Gaussian and salt & pepper noises. The performance among High-TV and some TV based denoising algorithms are also compared in terms of objective and subjective image assessment parameters for two test images and two true SAR images degraded by speckle noise.