A Technique for Image Denoising Based on Adaptive Piecewise Linear Filters and Automatic Parameter Tuning

Abstract

A novel technique for image denoising is presented. The proposed approach adopts multi-pass processing based on a two-parameter piecewise linear filter that aims at removing the noise without blurring the image structure. The accuracy of the filtering action is improved by an adaptive algorithm that avoids smoothing in presence of image details. A method for automatic parameter tuning based on image quality is also presented. Computer simulations show that the proposed approach yields very satisfactory results for images corrupted by Gaussian noise