Grounded Human-Object Interaction Video Generation on Real Scene 50 cases (test)
28.49VLM-QAInteractAvatar
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| InteractAvatar2026.02 | 28.49 | 91 | 13.5 | 79.4 | 66.5 | 28.9 | 5.51 | |
| HuMo2026.02 | 26.23 | 84 | 11.6 | 78.1 | 63.5 | 28.8 | 5.33 | |
| HY-Avatar2026.02 | 25.29 | 81.7 | 10.8 | 72.4 | 64.4 | 28.5 | 5.45 | |
| Wan-S2V2026.02 | 24.97 | 85.1 | 11.5 | 74 | 64.6 | 28.5 | 5.36 |