Human-Scene Interaction Motion Generation on Lingo (eval)
2.882FIDDeSeG
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DeSeGHierarchical Latent=true, Cross-Attention=true, Physical Regularization=true, Scene Representation=3D occupancy grid2026.07 | 2.882 | 5.865 | 2.541 | 0.562 | 0.207 | 0.921 | |
| DeSeGPhysical Regularization=false, Scene Representation=3D occupancy grid2026.07 | 2.983 | 5.822 | 2.524 | 0.514 | 0.367 | 1.393 | |
| LingoScene Representation=3D occupancy grid2026.07 | 3.093 | 5.832 | 2.239 | 0.437 | 0.392 | 1.229 | |
| DeSeGCross-Attention=false, Scene Representation=3D occupancy grid2026.07 | 3.113 | 5.636 | 2.721 | 0.457 | 0.239 | 0.946 | |
| DeSeGHierarchical Latent=false, Scene Representation=3D occupancy grid2026.07 | 3.279 | 5.461 | 2.439 | 0.43 | 0.241 | 0.938 | |
| TRUMANSScene Representation=3D occupancy grid2026.07 | 3.286 | 4.984 | 3.036 | 0.413 | 0.404 | 1.825 | |
| TeSMoScene Representation=2D floor map2026.07 | 3.558 | 5.328 | 2.956 | 0.392 | 0.278 | 0.915 |