Mixed Noise Removal of Images using Sparse Estimation

Natural image have different types of noises and mixed noise removal from the natural image is a very difficult process. One typical kind of method is Additive White Gaussian Noise (AWGN) coupled with Impulse Noise (IN). Numerous mixed noise removal approaches are based on detection method that tend to generate artificial products when mixed noise increased. In this paper, we propose an efficient method called weighted encoded with sparse non local regularization (WESNR) for mixed noise removal. In this method soft impulse pixel works on both impulse noise and additive white Gaussian noise. In additive the image sparsity prior and non local self similarity prior become as a same term and encoding framework. Experimental results show that the proposed WESNR method with new dictionary learning achieves leading mixed noise removal performance in terms of both quantitative measures and visual quality thus denoising technique performs better in terms of the PSNR.

Keywords:

Mixed noise removal , weighted encoding , nonlocal regularization , sparse estimation