Table 1.

Summary of methods that, to the best of our knowledge, have used a deep-learning-based approach for CMR reconstruction and which have been referred to in this article.

References	Application	Method/Network architecture	Training/Validation data
Hauptmann et al. (74)	Cine MRI	3D U-net applied in post-processing to reduce streaking artifacts	Single-coil retrospective/Single-coil prospective
Kofler et al. (75)	Cine MRI	2D U-net applied to the spatio-temporal domain in post-processing	Single-coil retrospective/Single-coil prospective
Schlemper et al. (76)	Cine MRI	End-to-end cascade of CNN regularization blocks and data-consistency blocks	Single-coil retrospective/Single-coil retrospective
Fuin et al. (77)	CMRA	End-to-end cascade of Multi-Scale VNN regularization blocks with data-consistency operators	Multi-coil retrospective/Multi-coil prospective
Biswas et al. (78)	Cine MRI	End-to-end cascade of CNN operators, an analytically defined SToRM prior, and conjugate gradient data consistency steps	Multi-coil retrospective/Multi-coil retrospective
Qin et al. (79)	Cine MRI	End-to-end cascade of recurrent CNN regularization blocks and data-consistency blocks	Single-coil retrospective/Single-coil retrospective
Akçakaya (80)	Myocardial T1 mapping	CNN for k-space interpolation	Scan-specific Autocalibrating Signal data
Wang et al. (81)	Cine MRI	A first CNN for k-space interpolation followed by a concatenated CNN network architecture for image dealiazing	Single-coil retrospective/Single-coil retrospective