Research map / 3D vision / Human pose estimation
Uncertainty-Aware Human Mesh Recovery From Video by Learning Part-Based 3D Dynamics
Gun-Hee Lee, Seong-Whan Lee
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 uncertainty-aware static feature u_t (ablation) | VIBE-based temporal encoder · 3DPW | PA-MPJPE: 56.8 -> 56.3 | Improved |
| add uncertainty-aware static feature u_t (to model with d_t) (ablation) | Full model · 3DPW | PA-MPJPE: 53.0 -> 52.2 | Improved |
| add uncertainty-aware static feature u_t (ablation) | Full model · 3DPW | acceleration error: 9.4 -> 6.8 | Improved |
| add dynamic optical-flow feature d_t (ablation) | Full model · 3DPW | acceleration error: 7.8 -> 6.8 | Improved |
| add dynamic optical-flow feature d_t (ablation) | Full model · 3DPW | PA-MPJPE: 56.3 -> 52.2 | Improved |
| add both u_t and d_t (ablation) | VIBE baseline · 3DPW | acceleration error: 9.8 -> 6.8 | Improved |
| add part-based 3D dynamics predictor (ablation) | Full model · 3DPW | acceleration error: 12.7 -> 6.8 | Improved |
| add part-based 3D dynamics predictor (ablation) | Full model · 3DPW | PA-MPJPE: 55.7 -> 52.2 | Improved |
| add dynamic feature d_t to model without dynamics predictor (ablation) | Part-based model · 3DPW | acceleration error: 21.6 -> 12.7 | Improved |
| add dynamic feature d_t to model without dynamics predictor (ablation) | Part-based model · 3DPW | PA-MPJPE: 56.1 -> 55.7 | Improved |
Closest papers by meaning
- Learning 3D Human Dynamics From Video (CVPR 2019 · similarity 0.92)
- HuMoR: 3D Human Motion Model for Robust Pose Estimation (ICCV 2021 · similarity 0.91)
- Self-Supervised 3D Human Mesh Recovery from a Single Image with Uncertainty-Aware Learning (AAAI 2024 · similarity 0.91)
- FactorizedHMR: A Hybrid Framework for Video Human Mesh Recovery (NeurIPS 2026 · similarity 0.89)
- VideoMDM: Towards 3D Human Motion Generation From 2D Supervision (NeurIPS 2026 · similarity 0.89)
- HumanMM: Global Human Motion Recovery from Multi-shot Videos (CVPR 2025 · similarity 0.89)
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.
