Fully 3D Unrolled Magnetic Resonance Fingerprinting Reconstruction via Staged Pretraining and Implicit Gridding
Yonatan Urman, Mark Nishimura, Daniel R. Abraham, Xiaozhi Cao, Kawin Setsompop · 2026
Read Fully 3D Unrolled Magnetic Resonance Fingerprinting Reconstruction via Staged Pretraining and Implicit Gridding by Yonatan Urman, Mark Nishimura, Daniel R. Abraham, Xiaozhi Cao, Kawin Setsompop (2026) 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/PMC13527266.txt.
Description
Urman and colleagues present a 2026 imaging method that rebuilds sharp three-dimensional scans from a fast magnetic resonance technique called fingerprinting. It measures tissue traits in one quick pass, but turning signals into sharp 3D images takes heavy computing, and common fixes blur at high speed-up. Their method, SPUR-iG, learns part of the rebuild and grids raw signals with learned kernels instead of slow transforms.
Full 3D training is slow, so the team trained it in stages, each lifting quality while keeping run times in reach. At one-millimeter detail, SPUR-iG beat both a standard rebuild and a recent hybrid method that mixes 2D and 3D steps. It also held up on scans from a different maker's machines, a hard test for learned systems. Faster, sharper fingerprinting could bring precise tissue measures into routine care.
Catalog record
| CREATOR | Yonatan Urman, Mark Nishimura, Daniel R. Abraham, Xiaozhi Cao, Kawin Setsompop |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 65 KB |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |