White-Matter Lesion Segmentation on MSLJ unseen (test)
74.2DSCLST-AI***
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| LST-AI***Inputs=T1W and FLAIR scans2026.03 | 74.2 | 70.3 | 61.7 | — | 12.1 | — | |
| nnU-Net SSMSU (Ours)Method=Self-Supervised Multi-Stage Unlearning2026.03 | 57.4 | 47.5 | 47.9 | 27.3 | 37.1 | 1.4 | |
| nnU-Net*Trainer=Predefined default trainer in nnU-Net2026.03 | 52.3 | 40.3 | 38.9 | 15.1 | 50.7 | 2.4 | |
| nnU-Net* Preprocessing**Trainer=Predefined default trainer in nnU-Net, Preprocessing=denoise and N4 bias correction2026.03 | 50.7 | 38.4 | 43.1 | 19.4 | 51.2 | 3 | |
| Dinsdale et al.2026.03 | 50.1 | 38.9 | 40.1 | 49.2 | 46.8 | 3.4 | |
| nnU-Net No DAData Augmentation=false2026.03 | 47.5 | 35.7 | 37.5 | 47.6 | 51.9 | 4.8 |