Research map / 3D vision / Human pose estimation
Reconstructing Groups of People with Hypergraph Relational Reasoning
Buzhen Huang, Jingyi Ju, Zhihao Li, Yangang Wang
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 hypergraph relational reasoning over individuals (ablation) | GigaCrowd mesh recovery · GigaCrowd | PA-PPDS: 70.20 -> 82.21 | Improved |
| use hypergraph reasoning instead of transformer reasoning (ablation) | GigaCrowd mesh recovery · GigaCrowd | PA-PPDS: 76.47 -> 82.21 | Improved |
| add more hypergraph scales (1 to 1,3 to 1,3,5) (ablation) | hypergraph network · GigaCrowd | PA-PPDS: 75.71 -> 78.11 -> 82.21 | Improved |
| add group size 11 as a fourth scale (ablation) | hypergraph-(1,3,5) · GigaCrowd | OKS: 77.31 -> 75.41 | Hurt |
| add group size 2 as a fourth scale (ablation) | hypergraph-(1,3,5) · GigaCrowd | PA-PPDS: 82.21 -> 82.61 | No clear effect |
| add crowd constraint loss (ablation) | hypergraph-(1,3,5) · GigaCrowd | PA-PPDS: 77.07 -> 82.21 | Improved |
| train with the proposed Panda pseudo ground-truth dataset (ablation) | Ours, BEV, CRMH · GigaCrowd | PA-PPDS: Ours 67.22 -> 82.21; BEV 55.41 -> 71.37; CRMH 52.16 -> 63.29 | Improved |
| add hypergraph reasoning compared with individual-only regression (ablation) | running speed · Panoptic | FPS: 26.4 -> 23.2 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PandaNet — PandaNet: Anchor-Based Single-Shot Multi-Person 3D Pose Estimation (CVPR 2020) · better on 1 of 1 shared result
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
- Learning To Estimate Robust 3D Human Mesh From In-the-Wild Crowded Scenes (CVPR 2022 · similarity 0.88)
- RSGNet: Relation based Skeleton Graph Network for Crowded Scenes Pose Estimation (AAAI 2021 · similarity 0.88)
- CrowdGaussian: Reconstructing High-Fidelity 3D Gaussians for Human Crowd from a Single Image (CVPR 2026 · similarity 0.87)
- MultiPly: Reconstruction of Multiple People from Monocular Video in the Wild (CVPR 2024 · similarity 0.86)
- Reconstructing People, Places, and Cameras (CVPR 2025 · similarity 0.85)
- COMPOSE: Hypergraph Cover Optimization for Multi-view 3D Human Pose Estimation (NeurIPS 2026 · similarity 0.85)
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.
