nnU-Net v2 — BraTS 2026 Brain Metastasis Segmentation
nnU-Net v2 (3D full-resolution) checkpoint trained on the BraTS 2026 Brain Metastasis Segmentation training set (fold 0, checkpoint_best.pth). Part of a three-model ensemble; see the full pipeline at NicoloPecco/MedNeXt_BraTS_Metastases.
Model Details
- Architecture: default nnU-Net residual encoder, 3D full-resolution
- Input: 4 MRI modalities — T1n, T1c, T2w, T2f
- Output: 5-class segmentation (background, NETC, SNFH, ET, RC)
- Trainer:
nnUNetTrainer - Checkpoint:
checkpoint_best.pth(fold 0)
Label Convention
| Label | Region |
|---|---|
| 0 | Background |
| 1 | NETC (Non-Enhancing Tumor Core) |
| 2 | SNFH (Surrounding Non-Enhancing FLAIR Hyperintensity) |
| 3 | ET (Enhancing Tumor) |
| 4 | RC (Resection Cavity) |
Usage
Inference code and full ensemble pipeline: GitHub repository
Install nnU-Net v2 following the instructions at MIC-DKFZ/nnUNet.
Related Models
Citation
@article{TODO}