Image denoising based on weighted regularized least square method

Noise in an image is caused due to various reasons. Removal of noise in an efficient way is a big challenge for researchers. In this paper, one dimensional signal denoising based on weighted regularized least square method is mapped to two dimensional image denoising. The proposed technique of image denoising based on least square is experimented on standard images sub-jected to different noises with varying noise levels. The effectiveness of the proposed method of image denoising is compared against the recently pro-posed sparse banded filter based image denoising technique. The comparison is done based on standard metric called peak signal to noise ratio(PSNR). The experimental result analysis shows that the proposed technique of least square image denoising outperforms the recently proposed sparse banded fil-ter based image denoising. The advantages of the proposed method are sim-ple computation and fast processing.

I. Introduction

The digital images are potentially degraded by noise through various sources such as background illuminations, image acquisition and transmission [4]. The image can be affected by many kind of noises, in which the respective noise level may vary from low level to high level. Depending upon the level of noise and the type of noise, the image denoising technique [9] can be applied. Image denoising is the process of improving the quality of the image by removing the noise content. The domains which require efficient noise removal are medical diagnosis [1], image segmentation, pattern recognition, image restoration, object recognition and video analysis [2].