Panoptic video scene graph generation on PSG4D-GTA
9.07R@20Ours
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| OursInput type=RGB-D videos2026.01 | 9.07 | 5.52 | 9.73 | 6.32 | 9.73 | 6.32 | |
| Motion-aware Contrastive LearningInput type=RGB-D videos2024.12 | 9.07 | 5.52 | 9.73 | 6.32 | 9.73 | 6.32 | |
| PSG4DFormerInput type=RGB-D videos2026.01 | 6.68 | 3.31 | 7.17 | 3.85 | 7.22 | 4.02 | |
| PSG4DFormerInput type=RGB-D videos2024.12 | 6.68 | 3.31 | 7.17 | 3.85 | 7.22 | 4.02 | |
| OursInput type=Point cloud videos2026.01 | 5.88 | 3.45 | 6.31 | 3.7 | 7.31 | 4.7 | |
| Motion-aware Contrastive LearningInput type=Point cloud videos2024.12 | 5.88 | 3.45 | 6.31 | 3.7 | 7.31 | 4.7 | |
| PSG4DFormerInput type=Point cloud videos2026.01 | 4.33 | 2.93 | 4.83 | — | 5.22 | 3.13 | |
| PSG4DFormerInput type=Point cloud videos2024.12 | 4.33 | 2.1 | 4.83 | 2.93 | 5.22 | 3.13 | |
| 3DSGGInput type=RGB-D videos2026.01 | 2.29 | 0.92 | 2.46 | 1.01 | 3.81 | 1.45 | |
| 3DSGGInput type=RGB-D videos2024.12 | 2.29 | 0.92 | 2.46 | 1.01 | 3.81 | 1.45 | |
| 3DSGGInput type=Point cloud videos2026.01 | 1.48 | 0.73 | 2.16 | 0.79 | 2.92 | 0.85 | |
| 3DSGGInput type=Point cloud videos2024.12 | 1.48 | 0.73 | 2.16 | 0.79 | 2.92 | 0.85 |