3D LiDAR Anomaly Segmentation on STU (test)
93.67AUROCLIDO
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LIDO2026.04 | 93.67 | 34.29 | 14.99 | |
| Mask4Former3D + Deep EnsembleBase Architecture=Mask4Former3D, Anomaly Detection Method=Deep Ensemble2026.04 | 86.74 | 58.05 | 5.17 | |
| Mask4Former3D + Void ClassifierBase Architecture=Mask4Former3D, Anomaly Detection Method=Void Classifier2026.04 | 85.99 | 78.6 | 3.92 | |
| Mask4Former3D + Max LogitBase Architecture=Mask4Former3D, Anomaly Detection Method=Max Logit2026.04 | 84.53 | 81.49 | 0.95 | |
| Mask4Former3D + RbABase Architecture=Mask4Former3D, Anomaly Detection Method=RbA2026.04 | 66.38 | 100 | 0.81 | |
| Mask4Former3D + MC DropoutBase Architecture=Mask4Former3D, Anomaly Detection Method=MC Dropout2026.04 | 61.51 | 82.37 | 0.11 |