Research map / 3D vision / Human pose estimation
DGG-HMR: Multi-Person Human Mesh Recovery with Depth-Guided Geometric Anchoring
Yanjie Li, Le Hui, Yali Peng, Shigang Liu
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 instance-level pelvis depth token concatenated with image tokens (ablation) | DGG-HMR baseline · AGORA val | MPJPE: 77.3 -> 74.3 | Improved |
| add instance-level depth token (ablation) | DGG-HMR baseline · AGORA val | PCDR0.2 (%): 89.3 -> 92.4 | Improved |
| add geometry-anchored refinement decoder on top of depth token (ablation) | DGG-HMR · AGORA val | MPJPE: 74.3 -> 69.0 | Improved |
| add geometry-anchored refinement decoder on top of depth token (ablation) | DGG-HMR · AGORA val | ADE (m): 2.20 -> 2.12 | Improved |
| use single-stage joint training instead of two-stage (depth pre-training then joint fine-tuning) (ablation) | DGG-HMR · AGORA val | MPJPE: 75.3 -> 69.0 | Improved |
| use single-stage joint training instead of two-stage (ablation) | DGG-HMR · AGORA val | PCDR0.2 (%): 93.9 -> 94.8 | Improved |
| fuse geometry priors into image features (feature-level concatenation) rather than injecting into decoder queries (ablation) | DGG-HMR · design discussion | depth ambiguity | No clear effect |
| increase input resolution from 672 to 896 (ablation) | DGG-HMR · 3DPW | PA-MPJPE: 40.5 -> 39.0 | Improved |
| add instance-aware depth estimator and geometry anchoring versus SAT-HMR | DGG-HMR (896) · MuPoTS-3D | PCK-All: 89.0 -> 86.2 | Hurt |
| add instance-aware depth estimator and geometry anchoring versus SAT-HMR | DGG-HMR (896) · 3DPW | MPJPE: 63.6 -> 61.1 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- SAT-HMR — SAT-HMR: Real-Time Multi-Person 3D Mesh Estimation via Scale-Adaptive Tokens (CVPR 2025) · better on 7 of 11 shared results
- AiOS (1333) — AiOS: All-in-One-Stage Expressive Human Pose and Shape Estimation (CVPR 2024) · better on 0 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
- ADHMR: Aligning Diffusion-based Human Mesh Recovery via Direct Preference Optimization (ICML 2025 · similarity 0.88)
- FactorizedHMR: A Hybrid Framework for Video Human Mesh Recovery (NeurIPS 2026 · similarity 0.88)
- SAT-HMR: Real-Time Multi-Person 3D Mesh Estimation via Scale-Adaptive Tokens (CVPR 2025 · similarity 0.88)
- GenHMR: Generative Human Mesh Recovery (AAAI 2025 · similarity 0.88)
- Incorporating Test-Time Optimization into Training with Dual Networks for Human Mesh Recovery (NeurIPS 2024 · similarity 0.88)
- PostureHMR: Posture Transformation for 3D Human Mesh Recovery (CVPR 2024 · 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.
