Coil-sketched unrolled networks for computationally-efficient deep MRI reconstruction

Julio Oscanoa, Batu Ozturkler, Siddharth Iyer, Zhitao Li, Christopher M. Sandino, Mert Pilancı, Daniel B. Ennis, Shreyas VasanawalaProceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition 2023 · DOI

Abstract

Deep unrolled networks can outperform conventional compressed sensing reconstruction. However, training unrolled networks has intensive memory and computational requirements, and is limited by GPU-memory constraints. We propose to use our previously developed “coil-sketching” algorithm to lower the computational burden of the data consistency step. Our method reduced memory usage and training time by 18% and 15% respectively with virtually no penalty on reconstruction accuracy when compared to a state-of-the-art unrolled network.
BibTeX
@inproceedings{oscanoa2023coil,
  author = {Julio Oscanoa and Batu Ozturkler and Siddharth Iyer and Zhitao Li and Christopher M. Sandino and Mert Pilancı and Daniel B. Ennis and Shreyas Vasanawala},
  title = {{Coil-sketched unrolled networks for computationally-efficient deep MRI reconstruction}},
  booktitle = {Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition},
  year = {2023},
  doi = {10.58530/2022/0947},
}

Loading related content...

Figure