WCSE 2020 SPRING ISBN: 978-981-14-4787-7
DOI: 10.18178/wcse.2020.02.015

Compressed Sensing Image De-noising Algorithm Based on L1-L2 Norm Regularization

Liu Ziming, Fang Changjie

Abstract— In this paper, we propose a compressed sensing image de-noising algorithm based on L1-L2 norm regularization. After the image is decomposed by the total variation spectral framework, L1 norm regularization is performed on the texture image, and L2 norm regularization is performed on the contour image, then the alternating direction method of multipliers (ADMM) is used for solution. The results of numerical experiment show that the proposed algorithm obtains higher peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) than the compared compressed sensing algorithm and the total variation algorithm, and can effectively maintain the contour information and texture information of the image when de-noising.

Index Terms— compressed sensing, regularization, ADMM, total variation.

Liu Ziming, Fang Changjie
Chongqing University of Posts and Telecommunications, CHINA

[Download]


Cite: Liu Ziming, Fang Changjie, "A Fear State Judgement System for Alleviating Fear of Heights Gradually in VR," Proceedings of 2020 the 10th International Workshop on Computer Science and Engineering (WCSE 2020), pp. 84-89, Yangon (Rangoon), Myanmar (Burma), February 26-28, 2020.