Full-body manipulation generation on FullBodyManipulation (test)
0.28FIDOurs (Spatial+Temporal)
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ours (Spatial+Temporal)Img Enc=DINOv22026.06 | 0.28 | 0.0843 | 0.0765 | 0.0745 | 0.365 | 0.582 | 0.726 | 0.823 | 0.633 | 0.677 | 0.43 | 0.53 | 5.69 | |
| Ours (Spatial)Img Enc=DINOv22026.06 | 0.3 | 0.1011 | 0.0891 | 0.0795 | 0.352 | 0.569 | 0.716 | 0.811 | 0.621 | 0.663 | 0.44 | 0.56 | 6.28 | |
| CHOIS+ImgImg Enc=CLIP2026.06 | 0.47 | 0.199 | 0.1683 | 0.1308 | 0.333 | 0.544 | 0.688 | 0.796 | 0.625 | 0.655 | 0.41 | 0.58 | 3.31 | |
| Ours (Temporal)Img Enc=DINOv22026.06 | 0.52 | 0.1907 | 0.1623 | 0.1256 | 0.34 | 0.565 | 0.714 | 0.805 | 0.669 | 0.697 | 0.39 | 0.62 | 6.37 | |
| CHOIS+ImgImg Enc=DINOv22026.06 | 0.63 | 0.1961 | 0.1648 | 0.1323 | 0.33 | 0.543 | 0.687 | 0.805 | 0.587 | 0.636 | 0.45 | 0.6 | 3.7 | |
| CHOIS2026.06 | 0.69 | — | — | — | 0.322 | 0.534 | 0.64 | 0.8 | 0.64 | 0.67 | 0.35 | 0.59 | 2.87 | |
| ViHOI*Img Enc=Qwen2.5-VL2026.06 | 0.71 | 0.2055 | 0.1761 | 0.1333 | 0.313 | 0.517 | 0.67 | 0.8 | 0.611 | 0.648 | 0.42 | 0.59 | 3.43 | |
| CHOIS+ImgImg Enc=Qwen2.5-VL2026.06 | 0.75 | 0.2064 | 0.174 | 0.1311 | 0.317 | 0.51 | 0.653 | 0.798 | 0.534 | 0.64 | 0.4 | 0.62 | 3.93 | |
| MDM2026.06 | 6.16 | — | — | — | — | — | 0.51 | 0.72 | 0.47 | 0.53 | 0.48 | 0.66 | 19.42 | |
| InterDiff2026.06 | 20.8 | — | — | — | — | — | 0.08 | 0.63 | 0.28 | 0.33 | 0.42 | 0.55 | 72.72 |