Image Denoising via Combination Anisotropic Diffusion and Bilateral Filtering

Abstract

In this paper we propose a new image denoising technique based on combination of nonlinear diffusion and bilateral filtering. The proposed algorithm uses Weighted Local Variance (WLV) of the residual image to determine the texture and fine details. These regions are denoised by bilateral filtering and then, are returned to the image denoised by P-M method. In fact, the WLV have a controlling role in amount of this combination. Denoising the residual image and adding it to the primary denoised image is the main novel idea of this paper. Experimental results confirm this image denoising approach is more efficient than nonlinear diffusion and the bilateral filtering while each algorithm accomplished separately.