TY - JOUR
T1 - Optimization-Inspired Compact Deep Compressive Sensing
AU - Zhang, Jian
AU - Zhao, Chen
AU - Gao, Wen
N1 - KAUST Repository Item: Exported on 2020-10-01
PY - 2020
Y1 - 2020
N2 - In order to improve CS performance of natural images, in this paper, we propose a novel framework to design an OPtimization-INspired Explicable deep Network, dubbed OPINE-Net, for adaptive sampling and recovery. Both orthogonal and binary constraints of sampling matrix are incorporated into OPINE-Net simultaneously. In particular, OPINE-Net is composed of three subnets: sampling subnet, initialization subnet and recovery subnet, and all the parameters in OPINE-Net (\eg sampling matrix, nonlinear transforms, shrinkage threshold, step size) are learned end-to-end, rather than hand-crafted. Moreover, considering the relationship among neighboring blocks, an enhanced version OPINE-${\rm{Net^{+}}}$ is developed, which allows image blocks to be sampled independently but reconstructed jointly to further enhance the performance. In addition, some interesting findings of learned sampling matrix are presented. Compared with existing state-of-the-art network-based CS methods, the proposed hardware-friendly OPINE-Nets not only achieve better performance but also require much fewer parameters and much less storage space, while maintaining a real-time computational speed.
AB - In order to improve CS performance of natural images, in this paper, we propose a novel framework to design an OPtimization-INspired Explicable deep Network, dubbed OPINE-Net, for adaptive sampling and recovery. Both orthogonal and binary constraints of sampling matrix are incorporated into OPINE-Net simultaneously. In particular, OPINE-Net is composed of three subnets: sampling subnet, initialization subnet and recovery subnet, and all the parameters in OPINE-Net (\eg sampling matrix, nonlinear transforms, shrinkage threshold, step size) are learned end-to-end, rather than hand-crafted. Moreover, considering the relationship among neighboring blocks, an enhanced version OPINE-${\rm{Net^{+}}}$ is developed, which allows image blocks to be sampled independently but reconstructed jointly to further enhance the performance. In addition, some interesting findings of learned sampling matrix are presented. Compared with existing state-of-the-art network-based CS methods, the proposed hardware-friendly OPINE-Nets not only achieve better performance but also require much fewer parameters and much less storage space, while maintaining a real-time computational speed.
UR - http://hdl.handle.net/10754/661850
UR - https://ieeexplore.ieee.org/document/9019857/
UR - http://www.scopus.com/inward/record.url?scp=85081300208&partnerID=8YFLogxK
U2 - 10.1109/JSTSP.2020.2977507
DO - 10.1109/JSTSP.2020.2977507
M3 - Article
SN - 1932-4553
SP - 1
EP - 1
JO - IEEE Journal of Selected Topics in Signal Processing
JF - IEEE Journal of Selected Topics in Signal Processing
ER -