From 100,000+ images to winning the first brain MRI foundation model challenges: Sharing lessons and models
Not on the map. Why?
Abstract (arXiv)
Developing Foundation Models for medical image analysis is essential to overcome the unique challenges of radiological tasks. The first challenges of this kind for 3D brain MRI, SSL3D and FOMO25, were held at MICCAI 2025. Our solution ranked first in tracks of both contests. It relies on a U-Net CNN architecture combined with strategies leveraging anatomical priors and neuroimaging domain knowledge. Notably, our models trained 1-2 orders of magnitude faster and were 10 times smaller than competing transformer-based approaches. Models are available here: https://github.com/jbanusco/BrainFM4Challenges.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 7 of its 10 references that are in the map. First posted on arXiv 19 Jan 2026.
Builds on
7 of 10 references.
References · 7
- 2023GPT-4 Technical Reportcited by 18,132
- 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
- 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
- 2024Revisiting MAE pre-training for 3D medical image segmentationcited by 33
- 2024An OpenMind for 3D medical vision self-supervised learningcited by 23
- 2024AMAES: Augmented Masked Autoencoder Pretraining on Public Brain MRI Data for 3D-Native Segmentationcited by 16
- 2025A large-scale heterogeneous 3D magnetic resonance brain imaging dataset for self-supervised learningcited by 7
Built on by
Citing papers per year: 2026: 1.
1 paper.