Preprint Review Version 1 Preserved in Portico This version is not peer-reviewed

Review of Matrix Rank Constraint Model for Impulse Interference Image Inpainting

Version 1 : Received: 30 October 2023 / Approved: 30 October 2023 / Online: 30 October 2023 (09:28:05 CET)

A peer-reviewed article of this Preprint also exists.

Ma, S.; Li, Z.; Chu, F.; Fang, S.; Yang, W.; Li, L. Review of Matrix Rank Constraint Model for Impulse Interference Image Inpainting. Electronics 2024, 13, 470. Ma, S.; Li, Z.; Chu, F.; Fang, S.; Yang, W.; Li, L. Review of Matrix Rank Constraint Model for Impulse Interference Image Inpainting. Electronics 2024, 13, 470.

Abstract

Camera malfunctions or loss of storage elements in imaging devices may lead to loss of important image information or random impulse noise interference. Low rank is one of the important prior information in image optimization processing. This paper uses different low-rank constraint models of the image matrices to recover the impulse interference satellite images. Firstly, an overview of image inpainting models based on nuclear norm, truncated nuclear norm, weighted nuclear norm, and matrix-factorization-based F-norm is presented, and corresponding optimiza-tion iterative algorithms are provided. Then, we conducted experiments under three types of pulse interference and provided visual and numerical results. Finally, it is concluded that the WSVT_ADMM method based on the weighted nuclear norm can obtain the best image inpainting results; The UV_ADMM method based on F norm of matrix factorization has the least computa-tion time and can be used for large-scale low-rank matrix computation; The WSVT_ADMM method based on weighted nuclear norm and the TSVT_ ADMM method based on truncated nuclear norms can significantly improve the repair effect compared to the nuclear norm-based methods such as SVT, SVP, and n_ADMM.

Keywords

Low rank; Optimization model; Optimization algorithm; Image inpainting

Subject

Computer Science and Mathematics, Computer Vision and Graphics

Comments (0)

We encourage comments and feedback from a broad range of readers. See criteria for comments and our Diversity statement.

Leave a public comment
Send a private comment to the author(s)
* All users must log in before leaving a comment
Views 0
Downloads 0
Comments 0


×
Alerts
Notify me about updates to this article or when a peer-reviewed version is published.
We use cookies on our website to ensure you get the best experience.
Read more about our cookies here.