Phase retrieval from incomplete data via weighted nuclear norm minimization
作者：
Highlights：
• We propose a new scheme for solving the phase retrieval problem with noisy measure- ments using group sparse representation via weighted nuclear norm minimization.
• We suggest to use the ℓ1−2 metricdelity term for reconstruction of images form incomplete magnitude measurements corrupted by both salt-and-pepper and random- valued impulse noise.
• We present closed-form solutions for sub-problems after decomposing the main opti- mization problem. We give the analytical justication of the closed-form solution for the weighted nuclear norm proximal mapping, so that we can prove the convergence within the group-based sparse representation framework.
摘要
•We propose a new scheme for solving the phase retrieval problem with noisy measure- ments using group sparse representation via weighted nuclear norm minimization.•We suggest to use the ℓ1−2 metricdelity term for reconstruction of images form incomplete magnitude measurements corrupted by both salt-and-pepper and random- valued impulse noise.•We present closed-form solutions for sub-problems after decomposing the main opti- mization problem. We give the analytical justication of the closed-form solution for the weighted nuclear norm proximal mapping, so that we can prove the convergence within the group-based sparse representation framework.
论文关键词：Phase retrieval,Partial magnitudes,Nuclear norm minimization,Impulse noise
论文评审过程：Received 14 November 2020, Revised 1 November 2021, Accepted 12 January 2022, Available online 14 January 2022, Version of Record 19 January 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.108537
