Research map / 3D vision / Human pose estimation
WHAC: World-grounded Humans and Cameras
Wanqi Yin, Zhongang Cai, Chen Wei, Fanzhou Wang, Ruisi Wang, Haiyi Mei, Weiye Xiao, Zhitao Yang, Qingping Sun, Atsushi Yamashita, Ziwei Liu, Lei Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add MotionVelocimeter to visual odometry for scale recovery (ablation) | WHAC · EMDB2 | WA-MPJPE: 376.0 -> 142.2 | Improved |
| add visual odometry to MotionVelocimeter for camera trajectory shape (ablation) | WHAC · EMDB2 | C-ATE: 134.1 -> 14.8 | Improved |
| add visual odometry to MotionVelocimeter (ablation) | WHAC · EMDB2 | C-AS: 1.10 -> 1.40 | Hurt |
| use GT or assumed focal length instead of dummy 5,000 (ablation) | WHAC · EMDB2 | WA-MPJPE: 604.6 -> 144.0 | Improved |
| use GT focal length instead of assumed focal length (ablation) | WHAC · EMDB2 | WA-MPJPE: 144.0 -> 142.2 | Improved |
| add temporal GRUs to SMPLer-X-B and depth from focal length | SMPLer-X-B · EMDB1 | T-PVE: 41298.0 -> 140.2 | Improved |
| add temporal modeling (GRUs) to camera-frame EHPS | SMPLer-X-B · EgoBody EgoSet | Accl.: 18.9 -> 11.6 | Improved |
| add temporal GRUs to SMPLer-X-B | SMPLer-X-B · WHAC-A-Mole (camera-frame) | PA-MPJPE: 76.7 -> 76.5 | No clear effect |
| add temporal GRUs to SMPLer-X-B | SMPLer-X-B · WHAC-A-Mole (camera-frame) | PVE-hand: 70.6 -> 77.7 | Hurt |
| use WHAC world-grounded pipeline over SMPLer-X-B + DPVO | SMPLer-X-B · WHAC-A-Mole | W-MPJPE: 842.3 -> 343.3 | Improved |
| use GT gyro instead of estimated camera rotation | WHAC · WHAC-A-Mole | W-MPJPE: 343.8 -> 343.3 | No clear effect |
Closest papers by meaning
- WHAM: Reconstructing World-grounded Humans with Accurate 3D Motion (CVPR 2024 · similarity 0.89)
- Reconstructing People, Places, and Cameras (CVPR 2025 · similarity 0.88)
- ODHSR: Online Dense 3D Reconstruction of Humans and Scenes from Monocular Videos (CVPR 2025 · similarity 0.88)
- BEDLAM2.0: Synthetic humans and cameras in motion (NeurIPS 2025 · similarity 0.87)
- HumanVid: Demystifying Training Data for Camera-controllable Human Image Animation (NeurIPS 2024 · similarity 0.87)
- Synergistic Global-space Camera and Human Reconstruction from Videos (CVPR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
