XtraLight-MedMamba for Classification of Neoplastic Tubular Adenomas
Aqsa Sultana, Rayan Afsar, Ahmed Rahu, MD · 2026
Read XtraLight-MedMamba for Classification of Neoplastic Tubular Adenomas by Aqsa Sultana, Rayan Afsar, Ahmed Rahu, MD (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-2602-04819.txt.
Description
Aqsa Sultana and 7 co-authors build a very small model to flag precancerous colon polyps in a 2026 medical imaging study. Sorting low-grade tubular adenomas by later cancer risk is hard because pathologists judge subtle patterns subjectively in whole-slide images.
Their model, XtraLight-MedMamba, blends a shallow convolutional feature extractor with parallel state-space blocks that model local texture within global context. An attention bridge module sharpens multi-scale features, and a special fixed classifier cuts parameters while improving generalization. The dataset grouped patients by whether colorectal cancer later developed. With only tens of thousands of parameters, the model reached high accuracy and beat larger transformer and Mamba rivals, suiting it to resource-constrained settings.
Catalog record
| CREATOR | Aqsa Sultana, Rayan Afsar, Ahmed Rahu, MD |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-04819