WCSE

A Shearlet-TV Regularization Approach for Image Denoising in the ADMM Framework

Abstract— The Shearlet transform is widely used in images denoising. However, the Shearlet based denoising method suffers from Gibbs artifacts because of the cut off in Fourier domain. In order to suppress the Gibbs artifacts, we propose a multi-constrained denoising model by using Shearlet transform and total variation. Considering that the proposed model is a multi-constrained problem, alternating direction method of multipliers is used to decouple the complex issue into some sub problems, which are easier to solve. Experiments are carried out to show the performance of proposed method. The experimental results indicate that the proposed method effectively relive the Gibbs artifacts of Shearlet transform.

Index Terms— shearlet, TV, ADMM, image denoising

Xingguo Liu, Yingpin Chen, Zhenming Pen
School of Optoelectronic Information, University of Electronic Science and Technology of China, CHINA
Xingguo Liu
Chongqing College of Electronic Engineering, CHINA
Juan Wu
College of Communication Engineering, Chongqing University, CHINA

[Download]

Cite: Xingguo Liu, Yingpin Chen, Zhenming Pen, Juan Wu, "A Shearlet-TV Regularization Approach for Image Denoising in the ADMM Framework," Proceedings of 2017 the 7th International Workshop on Computer Science and Engineering, pp. 283-287, Beijing, 25-27 June, 2017.