Research map / 3D vision / Human pose estimation
WorldPose: A World Cup Dataset for Global 3D Human Pose Estimation
Tianjian Jiang, Johsan Billingham, Sebastian Müksch, Juan J Zarate, Nicolas Evans, Martin R. Oswald, Marc Pollefeys, Otmar Hilliges, Manuel Kaufmann, Jie Song
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 bundle adjustment and SMPL fitting to the base multi-view pipeline (ablation) | WorldPose annotation pipeline · Vicon setup | G-MPJPE (mm): 83.5 -> 80.0 | Improved |
| add SMPL fitting and BA to base pipeline (ablation) | WorldPose annotation pipeline · Vicon setup | PA-MPJPE (mm): 70.8 -> 66.3 | Improved |
| bundle adjustment optimizing cameras w.r.t. field markings only (ablation) | WorldPose annotation pipeline · Vicon setup | G-MPJPE (mm): 83.5 -> 548.4 | Hurt |
| bundle adjustment on both field markings and player keypoints without SMPL (ablation) | WorldPose annotation pipeline · Vicon setup | G-MPJPE (mm): 83.5 -> 86.2 | No clear effect |
| refine broadcasting camera calibration with 3D player poses (ablation) | broadcast camera calibration · WorldPose broadcast clips | player reprojection alignment (qualitative) | Improved |
| replace DROID-SLAM cameras with ground-truth cameras in SLAHMR (ablation) | SLAHMR · WorldPose benchmark | G-MPJPE (mm): 8 334.1 -> 5 837.2 | Improved |
| remove HuMoR motion prior from SLAHMR (ablation) | SLAHMR · WorldPose benchmark | PA-MPJPE (mm): 163.9 -> 140.9 | Improved |
| align each player's trajectory individually (per-person) for GLAMR (ablation) | GLAMR · WorldPose benchmark | G-MPJPE (mm): 18 888.9 -> 3 749.7 | Improved |
| use GLAMR or SLAHMR global optimization instead of its initializer (HybrIK or 4DHuman) | global pose estimation · WorldPose benchmark | PA-MPJPE (mm): 78.8 -> 85.2 (GLAMR); 116.5 -> 163.9 (SLAHMR) | Hurt |
Closest papers by meaning
- CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wild (CVPR 2021 · similarity 0.86)
- TROPHIES: Temporal Reconstruction of Places, Humans, and Cameras from Multi-view Videos (CVPR 2026 · similarity 0.85)
- Lifting Motion to the 3D World via 2D Diffusion (CVPR 2025 · similarity 0.84)
- Reconstructing People, Places, and Cameras (CVPR 2025 · similarity 0.84)
- FreeMan: Towards Benchmarking 3D Human Pose Estimation under Real-World Conditions (CVPR 2024 · similarity 0.84)
- HumanMM: Global Human Motion Recovery from Multi-shot Videos (CVPR 2025 · similarity 0.83)
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.
