Table 1 Summary of different frameworks and algorithms for SCI reconstruction

optimization

TwIST, GAP-TV, DeSCI, GMM, KSVD

flexible, diverse quality, slow

[33, 45, 46, 50, 57]

deep learning

Tensor ADMM-Net,E2E-CNN, BIRNAT, MetaSCI, 3D-CNN, RevSCI

fast inference, high quality, inflexible, large GPU memory consumption, long training time, extensive training data

[35, 37, 51,52,53,54,55, 58]

plug-and-play

PnP-FFDNet, PnP-TV-FastDVDNet

flexible, moderate speed, moderate quality

[42, 47, 59]