BEHAVE is a full body human-object interaction dataset with multi-view RGBD frames and corresponding 3D SMPL and object fits along with the annotated contacts between them. Dataset contains ~15k frames at 5 locations with 8 subjects performing a wide range of interactions with 20 common objects.
Variants: BEHAVE
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| Contact Detection | CONTHO | Joint Reconstruction of 3D Human … | 2024-04-07 |
| 3D Object Reconstruction | CONTHO | Joint Reconstruction of 3D Human … | 2024-04-07 |
| Contact Detection | DECO | DECO: Dense Estimation of 3D … | 2023-09-26 |
| Contact Detection | BSTRO | Capturing and Inferring Dense Full-Body … | 2022-06-20 |
| 3D Object Reconstruction | CHORE | CHORE: Contact, Human and Object … | 2022-04-05 |
| Contact Detection | POSA | Populating 3D Scenes by Learning … | 2020-12-21 |
| 3D Object Reconstruction | PHOSA | Perceiving 3D Human-Object Spatial Arrangements … | 2020-07-30 |
Recent papers with results on this dataset: