Segmentation of Ischemic Stroke Lesions using Transfer Learning on Multi-sequence MRI
Rishiraj Paul Chowdhury, Tanjim Rahman · 2025
Read Segmentation of Ischemic Stroke Lesions using Transfer Learning on Multi-sequence MRI by Rishiraj Paul Chowdhury, Tanjim Rahman (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-2511-07281.txt.
Description
R P Chowdhury and T Rahman segment ischemic stroke lesions in brain MRI automatically, sparing experts a tedious and inconsistent manual task. MRI is sensitive to acute stroke, but outlining lesions by hand is slow and varies between observers. Their system works across T1-weighted, T2-weighted, diffusion-weighted, and FLAIR sequences, and is built around a Res-Unet design tested with and without pre-trained weights to isolate the benefit of transfer learning.
Validation used the ISLES 2015 brain stroke dataset, scoring three-dimensional volumes with Dice score and sensitivity. A majority voting classifier then merged the predictions made along each axis into one final segmentation. The combined system reached a Dice score of 80.5 percent and an accuracy of about seventy-four percent, which the authors present as evidence the transfer learning approach is effective.
Catalog record
| CREATOR | Rishiraj Paul Chowdhury, Tanjim Rahman |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 20 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-07281