White-Matter Lesion Segmentation on MSSEG unseen (test)
65.9DSCLST-AI***
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| LST-AI***Inputs=T1W and FLAIR scans2026.03 | 65.9 | 67.5 | 81.5 | — | 9.4 | — | |
| nnU-Net SSMSU (ours)Method=Self-Supervised Multi-Stage Unlearning2026.03 | 63.5 | 56.3 | 70 | 38.4 | 42 | 1.6 | |
| nnU-Net*Trainer=Predefined default trainer in nnU-Net2026.03 | 58.8 | 47.9 | 61.9 | 28.1 | 51.6 | 2.6 | |
| Dinsdale et al.2026.03 | 56.4 | 54.8 | 74 | 70.8 | 82.8 | 3 | |
| nnU-Net* Preprocessing**Trainer=Predefined default trainer in nnU-Net, Preprocessing=denoise and N4 bias correction2026.03 | 54.8 | 44.2 | 58.4 | 33.7 | 49.3 | 3.6 | |
| nnU-Net No DAData Augmentation=false2026.03 | 51.5 | 46.8 | 62.8 | 68.4 | 87.7 | 4.2 |