2019/03/23 by Yunyi Li, Fei Dai, Li, Yunyi +7
Engineering · Medicine · #FOS: Electrical engineering #Image and Video Processing (eess.IV) #Microwave Imaging and Scattering Analysis #Photoacoustic and Ultrasonic Imaging #Sparse and Compressive Sensing Techniques #Ultrasound Imaging and Elastography #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.1903.09787
openalex publication_date 2019/03/23 · openalex created_date 2022/07/29 · openalex updated_date 2026/07/28
Group-prior based regularization method has led to great successes in various\nimage processing tasks, which can usually be considered as a low-rank matrix\nminimization problem. As a widely used surrogate function of low-rank, the\nnuclear norm based convex surrogate usually lead to over-shrinking phenomena,\nsince the nuclear norm shrinks the rank components (singular value)\nsimultaneously. In this paper, we propose a novel Group-prior based nonconvex\nimage compressive sensing (CS) reconstruction framework via a family of\nnonconvex nuclear norms functions which contain common concave and\nmonotonically properties. To solve the resulting nonconvex nuclear norm\nminimization (NNM) problem, we develop a Group based iteratively\ndouble-reweighted nuclear norm algorithm (IDNN) via an alternating direction\nmethod of multipliers (ADMM) framework. Our proposed algorithm can convert the\nnonconvex nuclear norms optimization problem into a double-reweighted singular\nvalue thresholding (DSVT) problem. Extensive experiments demonstrate our\nproposed framework achieved favorable reconstruction performance compared with\ncurrent state-of-the-art convex methods.\n