# LightMedSeg-ISLES: Stroke Lesion Segmentation with 81x Fewer Parameters than nnU-Net

arXiv: https://arxiv.org/abs/2609.09634 · HTML: https://mlsanity.com/paper/2609.09634/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 13 references that are in the map.
- First posted on arXiv 9 Sep 2026.

## Builds on (most cited first, 7 of 7)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [FiLM: Visual Reasoning with a General Conditioning Layer](https://mlsanity.com/paper/1709.07871/) (2017)
- [GhostNet: More Features from Cheap Operations](https://mlsanity.com/paper/1911.11907/) (2019)
- [UNETR++: Delving into Efficient and Accurate 3D Medical Image Segmentation](https://mlsanity.com/paper/2212.04497/) (2022)
- [nnU-Net for Brain Tumor Segmentation](https://mlsanity.com/paper/2011.00848/) (2020)
- [How We Won the ISLES'24 Challenge by Preprocessing](https://mlsanity.com/paper/2505.18424/) (2025)
- [LightMedSeg: Lightweight 3D Medical Image Segmentation with Learned Spatial Anchors](https://mlsanity.com/paper/2603.07228/) (2026)

## Abstract (arXiv)

Large networks and ensembles often lead medical image segmentation challenges, but their storage and inference demands complicate deployment. We present LightMedSeg-ISLES, a 1.26-million-parameter pipeline for T1-weighted stroke lesion segmentation in ISLES'26. On a 146-case held-out cohort, flip test-time augmentation produces 0.618 mean Dice and 0.599 lesion-wise F1. A 102.35-million-parameter nnU-Net ResEnc-L produces 0.634 Dice and 0.544 lesion-wise F1 after size filtering. LightMedSeg therefore retains 97.5% of nnU-Net's Dice with 81.4× fewer parameters while improving lesion-wise F1 by 0.055. Its four-pass TTA operating point requires 4.7× fewer FLOPs per standardized patch than nnU-Net. It also slightly exceeds filtered UNETR++ and nnFormer. Longer training and stronger augmentation add 0.0358 Dice without increasing capacity, establishing a strong single-checkpoint alternative to much larger models.