A Reverse Mamba Attention Network for Pathological Liver Segmentation
1 st Jun Zeng, 2 nd Debesh Jha, 3 rd Ertugrul Aktas, 4 th Elif Keles, 5 th Alpay Medetalibeyoglu, 6 th Matthew Antalek, 7 th Robert Lewandowski, 8 th Daniela Ladner, 9 th Amir A. Borhani, 10 th Gorkem Durak, 11 th Ulas Bagci · 2025
Read A Reverse Mamba Attention Network for Pathological Liver Segmentation by 1 st Jun Zeng, 2 nd Debesh Jha, 3 rd Ertugrul Aktas, 4 th Elif Keles, 5 th Alpay Medetalibeyoglu, 6 th Matthew Antalek, 7 th Robert Lewandowski, 8 th Daniela Ladner, 9 th Amir A. Borhani, 10 th Gorkem Durak, 11 th Ulas Bagci (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-2502-18232.txt.
Description
Jun Zeng and co-authors present a vision model that outlines diseased livers in scans with state-of-the-art scores. Their 2025 paper builds on state space models, which read long sequences well but can blur fine local detail. A reverse attention module sharpens feature focus across scales in their design.
The model, RMA-Mamba, is tested on the hard task of pathological liver outlining. Tissue change in disease fools many standard methods there. On a new cirrhotic liver MRI set, the model posts top overlap and recall scores. On a cancerous liver CT set it also leads, with a Dice score near 93 percent. Code is public. The results argue the design travels well across scan types.
Catalog record
| CREATOR | 1 st Jun Zeng, 2 nd Debesh Jha, 3 rd Ertugrul Aktas, 4 th Elif Keles, 5 th Alpay Medetalibeyoglu, 6 th Matthew Antalek, 7 th Robert Lewandowski, 8 th Daniela Ladner, 9 th Amir A. Borhani, 10 th Gorkem Durak, 11 th Ulas Bagci |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-18232