3D Medical Image Segmentation on FLARE 2021 (test)
88.6DSCClean
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| CleanVictim Model=3D UNet2026.05 | 88.6 | 10.2 | — | — | |
| LSPVictim Model=3D UNet, Perturbation budget=4/2552026.05 | 88.5 | 17.5 | 43 | 99.7 | |
| SEPVictim Model=3D UNet, Perturbation budget=4/2552026.05 | 88.4 | 15.7 | 40.8 | 95.11 | |
| EMVictim Model=3D UNet, Perturbation budget=4/2552026.05 | 57.1 | 90.1 | 30 | 93.12 | |
| TAPVictim Model=3D UNet, Perturbation budget=4/2552026.05 | 45.7 | 136.7 | 37.7 | 90.84 | |
| PUEVictim Model=3D UNet, Perturbation budget=4/2552026.05 | 36.1 | 150.2 | 37.9 | 91.55 | |
| UMedVictim Model=3D UNet, Perturbation budget=4/2552026.05 | 26.3 | 107.3 | 48 | 99.73 | |
| VoxShieldVictim Model=3D UNet, Perturbation budget=4/2552026.05 | 6.8 | 172.4 | 49 | 99.8 |