A multi-dynamic low-rank deep image prior (ML-DIP) for real-time 3D cardiovascular MRI
Chong Chen, Marc Vornehm, Preethi Chandrasekaran, Muhammad A. Sultan, Syed M. Arshad, Yingmin Liu, Yuchi Han, Rizwan Ahmad · 2025
Read A multi-dynamic low-rank deep image prior (ML-DIP) for real-time 3D cardiovascular MRI by Chong Chen, Marc Vornehm, Preethi Chandrasekaran, Muhammad A. Sultan, Syed M. Arshad, Yingmin Liu, Yuchi Han, Rizwan Ahmad (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-19404.txt.
Description
Chong Chen and co-authors present a reconstruction framework for heart MRI in a 2025 arXiv paper. Their method, ML-DIP, rebuilds 3D real-time movies of the heart from heavily undersampled data. It needs no fully sampled training sets.
Separate neural networks model the image content and the motion fields. The pair is trained per scan, directly on the undersampled data. Tests covered a digital phantom with simulated extra beats, ten healthy volunteers, and twelve patients with such beats. Phantom scores were high even at two-minute scans. In people, heart function measures matched standard cine scans, and image quality beat a rival method.
Catalog record
| CREATOR | Chong Chen, Marc Vornehm, Preethi Chandrasekaran, Muhammad A. Sultan, Syed M. Arshad, Yingmin Liu, Yuchi Han, Rizwan Ahmad |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 53 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-19404