Uncertainty-aware Pointmap Prediction on ScanNet++/TUM RGB-D/KITTI Aggregate
0.4902AURCMCD (T=16)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MCD (T=16)Inference Type=Sampling, Training Time (h)=–2026.05 | 0.4902 | 0.2726 | 0.2644 | |
| MASt3RInference Type=Single-pass, Training Time (h)=–2026.05 | 0.4155 | 0.2053 | 0.4057 | |
| HeteroInference Type=Single-pass, Training Time (h)=9.42026.05 | 0.3973 | 0.1872 | 0.4183 | |
| Trust3RInference Type=Single-pass, Training Time (h)=10.52026.05 | 0.3861 | 0.1684 | 0.4898 | |
| DeepEns (K=5)Inference Type=Sampling, Training Time (h)=100.22026.05 | 0.2992 | 0.0916 | 0.4556 |