iPiano-Net: Nonconvex optimization inspired multi-scale reconstruction network for compressed sensing
作者：
Highlights：
• A nonconvex optimization iPiano algorithm inspired CS reconstruction network is proposed.
• A deep learning powered solver is set up to handle nonconvex optimization issues.
• Only a single model can handle CS reconstruction with several measurement ratios even unseen ones.
• High quality and fast reconstruction can be acquired on two benchmark datasets.
• The proposed network is robust to different levels of noise.
摘要
•A nonconvex optimization iPiano algorithm inspired CS reconstruction network is proposed.•A deep learning powered solver is set up to handle nonconvex optimization issues.•Only a single model can handle CS reconstruction with several measurement ratios even unseen ones.•High quality and fast reconstruction can be acquired on two benchmark datasets.•The proposed network is robust to different levels of noise.
论文关键词：iPiano algorithm,Compressed sensing,Convolutional neural network,Deep learning
论文评审过程：Received 16 October 2019, Revised 16 May 2020, Accepted 26 August 2020, Available online 1 September 2020, Version of Record 14 September 2020.
论文官网地址：https://doi.org/10.1016/j.image.2020.115989
