Character-video-driven human reenactment on 160 real-world motion videos (test)
6.619MSCAnyAct
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AnyActBase motion generation model=MoMask++2026.05 | 6.619 | 6.763 | 6.7 | 9.056 | |
| EchoMotionCategory=Video2SMPL(X), Video diffusion model=Wan2.2-5B2026.05 | 6.213 | 6.313 | 6.294 | 8.325 | |
| VLM + HY-MotionCategory=VLM+T2M, Base motion generation model=HY-Motion, VLM Prompt Generator=Qwen3.5-27B2026.05 | 5.794 | 6.025 | 5.881 | 9.181 | |
| GENMOCategory=Video2SMPL(X)2026.05 | 5.563 | 5.775 | 5.644 | 7.644 | |
| VLM + MoMask++Category=VLM+T2M, Base motion generation model=MoMask++, VLM Prompt Generator=Qwen3.5-27B2026.05 | 5.544 | 5.663 | 5.525 | 9.033 |