English

Deep Low-rank plus Sparse Network for Dynamic MR Imaging

Image and Video Processing 2021-07-21 v3 Computer Vision and Pattern Recognition Machine Learning Signal Processing

Abstract

In dynamic magnetic resonance (MR) imaging, low-rank plus sparse (L+S) decomposition, or robust principal component analysis (PCA), has achieved stunning performance. However, the selection of the parameters of L+S is empirical, and the acceleration rate is limited, which are common failings of iterative compressed sensing MR imaging (CS-MRI) reconstruction methods. Many deep learning approaches have been proposed to address these issues, but few of them use a low-rank prior. In this paper, a model-based low-rank plus sparse network, dubbed L+S-Net, is proposed for dynamic MR reconstruction. In particular, we use an alternating linearized minimization method to solve the optimization problem with low-rank and sparse regularization. Learned soft singular value thresholding is introduced to ensure the clear separation of the L component and S component. Then, the iterative steps are unrolled into a network in which the regularization parameters are learnable. We prove that the proposed L+S-Net achieves global convergence under two standard assumptions. Experiments on retrospective and prospective cardiac cine datasets show that the proposed model outperforms state-of-the-art CS and existing deep learning methods and has great potential for extremely high acceleration factors (up to 24x).

Keywords

Cite

@article{arxiv.2010.13677,
  title  = {Deep Low-rank plus Sparse Network for Dynamic MR Imaging},
  author = {Wenqi Huang and Ziwen Ke and Zhuo-Xu Cui and Jing Cheng and Zhilang Qiu and Sen Jia and Leslie Ying and Yanjie Zhu and Dong Liang},
  journal= {arXiv preprint arXiv:2010.13677},
  year   = {2021}
}
R2 v1 2026-06-23T19:39:30.384Z