Lite ENSAM: a lightweight cancer segmentation model for 3D Computed Tomography
Agnar Martin Bjørnstad, Elias Stenhede, Arian Ranjbar · 2025
Read Lite ENSAM: a lightweight cancer segmentation model for 3D Computed Tomography by Agnar Martin Bjørnstad, Elias Stenhede, Arian Ranjbar (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-2511-01600.txt.
Description
Agnar Martin Bjørnstad, Elias Stenhede, and Arian Ranjbar present a lightweight model for outlining tumors in three-dimensional CT scans. Measuring whole tumor volume tracks treatment response better than the standard RECIST approach, which uses one diameter in one plane, but hand-drawing volumes is too slow for routine care.
Their model, Lite ENSAM, adapts the ENSAM architecture for efficiency and learns from scans annotated only with RECIST measurements. It was entered in the MICCAI FLARE 2025 pan-cancer segmentation task, where it scored a Dice similarity of about 60.7 percent and a normalized surface Dice of about 63.6 percent on the hidden test set. On public validation data it ran in about 14.4 seconds per scan on a plain CPU.
Catalog record
| CREATOR | Agnar Martin Bjørnstad, Elias Stenhede, Arian Ranjbar |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 17 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-01600