3D Object Articulation Prediction on MultiScan
70.3mIoUUNITE
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| UNITEInput Modality=RGB images, Training configuration=multi-task2025.12 | 70.3 | 6 | 29.6 | 10 | 3.7 | 12.3 | 6.9 | |
| S2MInput Modality=Point Cloud (reconstructed meshes)2025.12 | 69.2 | 6.7 | 15.7 | 9.4 | 4.5 | 10.4 | 6.3 | |
| UNITEInput Modality=RGB images, Training configuration=task-only2025.12 | 68.3 | 5.2 | 23.4 | 7.6 | 1.8 | 9.2 | 4.9 | |
| OPDPNInput Modality=Point Cloud (reconstructed meshes)2025.12 | 53.8 | 1.6 | 3.1 | 2.1 | 1.3 | 2.6 | 1.7 |