Research map / 3D vision / Human pose estimation
Affordance-Guided Diffusion Prior for 3D Hand Reconstruction
Naru Suzuki, Takehiko Ohkawa, Tatsuro Banno, Jihyun Lee, Ryosuke Furuta, Yoichi Sato
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 |
|---|---|---|---|
| Refine WiLoR with affordance-guided diffusion prior | WiLoR · HOGraspNet | PA-MPJPE avg: 8.22 -> 7.53 | Improved |
| Refine WiLoR with affordance prior, high occlusion subset | WiLoR · HOGraspNet | PA-MPJPE high occlusion: 12.41 -> 9.37 | Improved |
| Refine Hamba with affordance prior | Hamba · HOGraspNet | PA-MPJPE avg: 8.46 -> 7.85 | Improved |
| Refine WiLoR with affordance prior | WiLoR · HO3D | PA-MPJPE avg: 7.43 -> 7.37 | No clear effect |
| Apply InterHandGen diffusion prior without context | WiLoR · HOGraspNet | PA-MPJPE avg: 8.22 -> 8.99 | Hurt |
| Add simple VLM captions as text condition (ablation) | AffHandGen · HOGraspNet | PA-MPJPE avg: 9.28 -> 8.67 (no RGB); 7.90 -> 7.76 (RGB) | No clear effect |
| Add affordance descriptions as text condition (ablation) | AffHandGen · HOGraspNet | PA-MPJPE avg: 7.90 -> 7.53 | Improved |
| Add RGB image features as condition (ablation) | AffHandGen · HOGraspNet | PA-MPJPE avg: 8.43 -> 7.53 | Improved |
| Use ground-truth grasp instead of predicted grasp (ablation) | AffHandGen · HOGraspNet | PA-MPJPE avg: 7.53 -> 7.51 | No clear effect |
| Add Mistral summarization of Qwen keywords (ablation) | AffHandGen · HOGraspNet | PA-MPJPE avg: 7.59 -> 7.53 | Improved |
| Add 2D keypoint guidance (ablation) | AffHandGen refinement · HOGraspNet | PA-MPJPE avg: 7.65 -> 7.53 | Improved |
| Use HaMeR+DistilBERT encoders instead of CLIP (ablation) | AffHandGen · HOGraspNet | PA-MPJPE avg: 11.36 -> 7.53 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Hamba — Hamba: Single-view 3D Hand Reconstruction with Graph-guided Bi-Scanning Mamba (NeurIPS 2024) · better on 2 of 2 shared results
- HandOccNet — HandOccNet: Occlusion-Robust 3D Hand Mesh Estimation Network (CVPR 2022) · better on 2 of 2 shared results
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
- G-HOP: Generative Hand-Object Prior for Interaction Reconstruction and Grasp Synthesis (CVPR 2024 · similarity 0.91)
- HandBooster: Boosting 3D Hand-Mesh Reconstruction by Conditional Synthesis and Sampling of Hand-Object Interactions (CVPR 2024 · similarity 0.89)
- HOIDiffusion: Generating Realistic 3D Hand-Object Interaction Data (CVPR 2024 · similarity 0.88)
- Single-view Image to Novel-view Generation for Hand-Object Interactions (AAAI 2025 · similarity 0.88)
- InterHandGen: Two-Hand Interaction Generation via Cascaded Reverse Diffusion (CVPR 2024 · similarity 0.88)
- The Surprising Effectiveness of Video Diffusion Models for Hand Motion Reconstruction (NeurIPS 2026 · similarity 0.88)
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.
