Anomaly Segmentation on SemanticKITTI OoD (Single)
93.36AUROCLIDO
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LIDOsplit=Single2026.04 | 93.36 | 31.19 | 10.6 | |
| LIDOAnomaly Detection Method=LIDO2026.04 | 93.36 | 31.19 | 10.6 | |
| Mask4Former3D + Deep Ensemblesplit=Single2026.04 | 92.87 | 27.69 | 6.2 | |
| Mask4Former3D + Deep EnsembleBackbone=Mask4Former3D, Anomaly Detection Method=Deep Ensemble2026.04 | 92.87 | 27.69 | 6.2 | |
| MinkowskiNet + Deep EnsembleBackbone=MinkowskiNet, Anomaly Detection Method=Deep Ensemble2026.04 | 91.64 | 39.96 | 4.78 | |
| MinkowskiNet + Max LogitBackbone=MinkowskiNet, Anomaly Detection Method=Max Logit2026.04 | 88.7 | 52.26 | 2.49 | |
| MinkowskiNet + MC DropoutBackbone=MinkowskiNet, Anomaly Detection Method=MC Dropout2026.04 | 87.49 | 50.75 | 3.04 | |
| Mask4Former3D + Max Logitsplit=Single2026.04 | 76.49 | 99.63 | 4.84 | |
| Mask4Former3D + Max LogitBackbone=Mask4Former3D, Anomaly Detection Method=Max Logit2026.04 | 76.49 | 99.63 | 4.84 | |
| Mask4Former3D + RbAsplit=Single2026.04 | 59.68 | 100 | 4.56 | |
| Mask4Former3D + RbABackbone=Mask4Former3D, Anomaly Detection Method=RbA2026.04 | 59.68 | 100 | 4.56 | |
| MinkowskiNet + Void ClassifierBackbone=MinkowskiNet, Anomaly Detection Method=Void Classifier2026.04 | 47.74 | 87.02 | 0.22 | |
| MinkowskiNet + RbABackbone=MinkowskiNet, Anomaly Detection Method=RbA2026.04 | 40.38 | 95.45 | 0.19 |