Research map / 3D vision / Human pose estimation
MUG: Multi-human Graph Network for 3D Mesh Reconstruction from 2D Pose
Chenyan Wu, Yandong Li, Xianfeng Tang, Zhe Huang, James Z. 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 root-joint inter-human edges (1 per pair) (ablation) | MUG · RH (not fine-tuned) | PCDR0.2: 54.80 -> 57.37 | Improved |
| add root-joint inter-human edges (1 per pair) (ablation) | MUG · RH (fine-tuned) | PCDR0.2: 67.95 -> 69.87 | Improved |
| increase inter-human edges from 1 to 3 per pair (ablation) | MUG · RH (fine-tuned) | PCDR0.2: 69.87 -> 69.06 | Hurt |
| increase inter-human edges from 0 to 3 per pair (ablation) | MUG · 3DPW | MPVPE: 101.4 -> 102.4 | Hurt |
| use heterogeneous joint and mesh graph instead of mesh nodes only (ablation) | MUG · 3DPW | MPVPE: 103.9 -> 101.4 | Improved |
| use heterogeneous graph instead of joint-to-mesh upsampling (ablation) | MUG · 3DPW | MPVPE: 104.2 -> 101.4 | Improved |
| add absolute pose feature to normalized pose (ablation) | MUG · RH (not fine-tuned) | PCDR0.2: 34.8 -> 52.8 | Improved |
| add absolute pose feature to normalized pose (ablation) | MUG · 3DPW | MPVPE: 107.1 -> 105.0 | Improved |
| add full-pose and template joint features (all four feature components) (ablation) | MUG · 3DPW | MPVPE: 105.0 -> 101.4 | Improved |
| use NeuralAnnot instead of SMPLify pseudo-ground-truth (ablation) | MUG · 3DPW | MPVPE: 103.1 -> 101.4 | Improved |
| use ViTPose instead of HRNet 2D poses (ablation) | MUG · 3DPW | MPVPE: 103.9 -> 101.4 | Improved |
| add 16 px keypoint jitter at test time (ablation) | MUG · 3DPW | MPVPE: 101.4 -> 108.2 | Hurt |
Closest papers by meaning
- Multi-HMR: Multi-Person Whole-Body Human Mesh Recovery in a Single Shot (ECCV 2024 · similarity 0.87)
- Recovering 3D Human Mesh From Monocular Images: A Survey (TPAMI 2023 · similarity 0.87)
- Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Pose (ECCV 2020 · similarity 0.87)
- Self-Supervised 3D Human Mesh Recovery from a Single Image with Uncertainty-Aware Learning (AAAI 2024 · similarity 0.86)
- MultiGO: Towards Multi-level Geometry Learning for Monocular 3D Textured Human Reconstruction (CVPR 2025 · similarity 0.85)
- TokenHMR: Advancing Human Mesh Recovery with a Tokenized Pose Representation (CVPR 2024 · 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.
