Research map / 3D vision / Human pose estimation
Multi-THuMBS: Multi-person Tracking of 3D Human Meshes Beyond Video Shots
Jeongwan On, Muhammad Salman Ali, Muneeb Khan, Sunwoo Park, Inwoong Moon, Hyung Jin Chang, Jaekwang Kim, Seong Jong Ha, Seungryul Baek
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 mesh-to-point-cloud alignment in shared 3D space (ablation) | Multi-THuMBS · EgoHumans | W-MPJPE: 882.7 -> 278.8 | Improved |
| add three-stage progressive optimization (ablation) | Multi-THuMBS · EgoHumans | W-MPJPE: 491.9 -> 278.8 | Improved |
| add camera pose optimization (ablation) | Multi-THuMBS · EgoHumans | ATE: 1.40 -> 0.77 | Improved |
| add post-processing smoothing and cross-camera loss (ablation) | Multi-THuMBS · EgoHumans | Accel: 47.9 -> 27.3 | Improved |
| add appearance and pose cues to 3D-distance Re-ID (ablation) | Multi-THuMBS · EgoHumans | identity switches: 1.66 -> 0.97 | Improved |
| add appearance and pose cues to 3D-distance Re-ID (ablation) | Multi-THuMBS · EgoBody | identity switches: 0.00 -> 0.00 | No clear effect |
| use shared 3D space Re-ID instead of appearance-only KPR | Multi-THuMBS · EgoHumans | identity switches: 2.54 -> 0.97 | Improved |
| use VGGT shared space and propagate vs PromptHMR | Multi-THuMBS · EgoBody | W-MPJPE: 1228.1 -> 99.2 | Improved |
| multi-shot method vs PromptHMR on in-the-wild videos | Multi-THuMBS · AVA and sitcom videos | cross-shot PCK: 62.7 -> 90.7 | Improved |
| multi-shot method vs HSfM-dagger MPJPE on Harmony4D | Multi-THuMBS · Harmony4D | MPVPE: 257.6 (HSfM) vs 278.3 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- PromptHMR — PromptHMR: Promptable Human Mesh Recovery (CVPR 2025) · better on 4 of 4 shared results
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- CoMotion: Concurrent Multi-person 3D Motion (ICLR 2025 · similarity 0.88)
- HumanMM: Global Human Motion Recovery from Multi-shot Videos (CVPR 2025 · similarity 0.88)
- Reconstructing People, Places, and Cameras (CVPR 2025 · similarity 0.87)
- MultiPhys: Multi-Person Physics-aware 3D Motion Estimation (CVPR 2024 · similarity 0.87)
- Harmony4D: A Video Dataset for In-The-Wild Close Human Interactions (NeurIPS 2024 · similarity 0.87)
- HOT3D: Hand and Object Tracking in 3D from Egocentric Multi-View Videos (CVPR 2025 · 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.
