TuLaBM: Tumor-Biased Latent Bridge Matching for Contrast-Enhanced MRI Synthesis
Atharva Rege 1,2, Adinath Madhavrao Dukre 1, Numan Balci 3, Dwarikanath Mahapatra 4, Imran Razzak 1 · 2026
Read TuLaBM: Tumor-Biased Latent Bridge Matching for Contrast-Enhanced MRI Synthesis by Atharva Rege 1,2, Adinath Madhavrao Dukre 1, Numan Balci 3, Dwarikanath Mahapatra 4, Imran Razzak 1 (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/arxiv-2603-19386.txt.
Description
Rege and co-authors synthesize contrast-enhanced MRI scans from plain scans, skipping contrast dye. Dye adds cost and safety risk, yet tumors show up best with it. Their TuLaBM moves between scan types in a learned latent space via a Brownian bridge. A tumor-focused attention step boosts features in the tumor zone during the trip. A boundary loss keeps tumor edges sharp in the output. Tests on brain and liver data beat top baselines on image and tumor scores. The latent design also cuts compute and run time.
Catalog record
| CREATOR | Atharva Rege 1,2, Adinath Madhavrao Dukre 1, Numan Balci 3, Dwarikanath Mahapatra 4, Imran Razzak 1 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21221 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.19386