Image Reconstruction: From Compressed Sensing to Machine Learning
Thomas Küstner1
1University Hospital Tuebingen, Tuebingen, Germany

Synopsis

Keywords: Image acquisition: Reconstruction, Image acquisition: Machine learning, Cardiovascular: Cardiovascular

Cardiovascular MR (CMR) is a versatile non-invasive imaging that provides a comprehensive assessment of cardiac function and anatomy in a single examination. CMR plays a major role in the diagnosis and management of cardiovascular disease. When setting up and optimizing a clinical CMR protocol, the inherent trade-off between spatial and temporal resolution, scan time and signal-to-noise ratio (SNR) must be taken into consideration. Several approaches have been proposed to speed up CMR, including parallel imaging, k-t accelerated imaging, or pseudo-random sub-Nyquist sampling. An overview of compressed sensing, low-rank and deep learning reconstructions is presented for these accelerated scans.

Acknowledgements

No acknowledgement found.

References

No reference found.
Proc. Intl. Soc. Mag. Reson. Med. 31 (2023)