An Improved Algorithm for Anisotropic Diffusion Smoothing Noise

Abstract

Smoothing noise plays a very important role in digital image processing.The anisotropic diffusion is a selective smoothing technique that effectively employs intra-region smoothing without limitation and inhibits inter-region smoothing.In order to remove noise and preserve edges effectively,a more effective and adaptive diffusion algorithm is proposed in this paper.The algorithm is based on anisotropic diffusion model and the statistical properties of image with noise.The value of parameter can be adaptively selected by the proposed algorithm based on the different signal-to-noise ratio.It has been shown from the experiments that the improved scheme has superiority capability over the Perona-Malik scheme and it is a robust anisotropic diffusion.