Research map / 3D vision / 3D generation
DiffProxy: Multi-View Human Mesh Recovery via Diffusion-Generated Dense Proxies
renke wang, zhenyu zhang, Ying Tai, Jun Li, Jian Yang
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 hand refinement and test-time scaling (ablation) | DiffProxy · RICH | MPVPE: 46.5 -> 35.1 | Improved |
| add hand refinement (ablation) | DiffProxy · MoYo | MPVPE: 44.9 -> 39.6 | Improved |
| add test-time scaling (K=3) (ablation) | DiffProxy · MoYo | MPVPE: 39.6 -> 38.8 | Improved |
| add DINOv2 features (ablation) | DiffProxy · BEHAVE | MPVPE: 60.2 -> 39.9 | Improved |
| add T2I-Adapter (ablation) | DiffProxy · BEHAVE | MPVPE: 48.4 -> 39.9 | Improved |
| add text attention module (ablation) | DiffProxy · BEHAVE | MPVPE: 43.3 -> 39.9 | Improved |
| add epipolar attention module (ablation) | DiffProxy · BEHAVE | MPVPE: 41.9 -> 39.9 | Improved |
| add cross-modal attention module (ablation) | DiffProxy · BEHAVE | MPVPE: 41.9 -> 39.9 | Improved |
| use ground-truth cameras instead of VGGT-predicted cameras (ablation) | DiffProxy · BEHAVE | PA-MPVPE: 37.3 -> 31.3 | Improved |
| use diffusion proxy generator instead of fine-tuned DensePose (ablation) | DiffProxy · MoYo | MPVPE: 111.4 -> 38.8 | Improved |
| fine-tune DensePose on synthetic data (60 epochs) (ablation) | DensePose · MoYo | MPVPE: 127.8 -> 111.4 | Improved |
| use dense proxies instead of sparse keypoints (EasyMoCap) | fitting pipeline · RICH | MPVPE: 60.3 -> 35.1 | Improved |
Closest papers by meaning
- DPMesh: Exploiting Diffusion Prior for Occluded Human Mesh Recovery (CVPR 2024 · similarity 0.90)
- GenHMR: Generative Human Mesh Recovery (AAAI 2025 · similarity 0.89)
- FactorizedHMR: A Hybrid Framework for Video Human Mesh Recovery (NeurIPS 2026 · similarity 0.89)
- ADHMR: Aligning Diffusion-based Human Mesh Recovery via Direct Preference Optimization (ICML 2025 · similarity 0.88)
- DiffHuman: Probabilistic Photorealistic 3D Reconstruction of Humans (CVPR 2024 · similarity 0.88)
- DiSRT-In-Bed: Diffusion-Based Sim-to-Real Transfer Framework for In-Bed Human Mesh Recovery (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.
