Corrected rank residual constraint model for image denoising. Issue 13 (13th July 2022)

• Record Type: Journal Article
• Authors: Wu, Di; Zhang, Tao
• Is Part Of: IET image processing. Volume 16:Issue 13(2022)
• Page Start: 3659
• Page End: 3668
• Publication Date: 2022-07-13
• Subjects: Image processing -- Periodicals; 621.36705
• DOI: 10.1049/ipr2.12583
• Languages: English
• ISSNs: 1751-9659

Abstract

Abstract: In this paper, a novel image denoising model is proposed, named Corrected Rank Residual Constraint (CRRC). To overcome the drawback that L 1 norm penalty yields biased estimators and cannot achieve the best estimation performance, the proposed CRRC incorporates the adaptive correction term with L 1 minimization to improve the sparsity of the L 1 norm data fidelity term enforced on the rank residual. The existence and uniqueness of solutions of the proposed model are also studied, and the optimization method for solving the model is given. The studies show that the adaptive correction term can not only improve the sparsity of the rank residual but also overcome the over‐shrinkage of large singular values. Experimental results demonstrate that the proposed CRRC model outperforms many state‐of‐the‐art methods in both objective and perceptual quality.