Research map / 3D vision / Human pose estimation
Decoupled Iterative Refinement Framework for Interacting Hands Reconstruction from a Single RGB Image
Pengfei Ren, Chao Wen, Xiaozheng Zheng, Zhou Xue, Haifeng Sun, Qi Qi, Jingyu Wang, Jianxin Liao
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 iterative pixel-aligned refinement (ablation) | basic MANO regression · InterHand2.6M | MPJPE (mm): 12.44 -> 11.30 | Improved |
| add GCN for intra-hand joint interaction (ablation) | iterative refinement · InterHand2.6M | MPJPE (mm): 11.30 -> 10.94 | Improved |
| add transformer for inter-hand joint interaction (ablation) | iterative refinement · InterHand2.6M | MPJPE (mm): 11.30 -> 10.98 | Improved |
| combine GCN and transformer (ablation) | iterative refinement · InterHand2.6M | MPJPE (mm): 11.30 -> 10.73 | Improved |
| multi-plane feature projecting instead of single-plane (ablation) | DIR · InterHand2.6M | MPJPE (mm): 10.68 -> 10.49 | Improved |
| multi-plane feature projecting instead of heatmap projection (ablation) | DIR · InterHand2.6M | MPJPE (mm): 10.65 -> 10.49 | Improved |
| increase refinement stages from two to three (ablation) | DIR · InterHand2.6M | MPJPE (mm): 10.49 -> 10.44 | Improved |
| larger learning rate (1e-4 to 3e-4) in basic model (ablation) | basic regression model · InterHand2.6M | MPJPE (mm): 12.79 -> 12.44 | Improved |
| attention split, smooth L1 loss, motion blur augmentation in basic model (ablation) | basic regression model · InterHand2.6M | MPJPE (mm): 12.53 -> 12.44 | Improved |
| joint-node interaction with GCN+transformer vs IntagHand dense vertex interaction | DIR · InterHand2.6M | MPJPE (mm): 12.40 -> 10.49 | Improved |
| joint-based spatial modeling for hand collisions | DIR · InterHand2.6M | Interpenetration volume (cm^3): 4.37 -> 4.14 | Improved |
Closest papers by meaning
- Interacting Attention Graph for Single Image Two-Hand Reconstruction (CVPR 2022 · similarity 0.92)
- HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video (CVPR 2024 · similarity 0.89)
- BIGS: Bimanual Category-agnostic Interaction Reconstruction from Monocular Videos via 3D Gaussian Splatting (CVPR 2025 · similarity 0.88)
- SiMA-Hand: Boosting 3D Hand-Mesh Reconstruction by Single-to-Multi-View Adaptation (AAAI 2024 · similarity 0.88)
- Fine-Grained Multi-View Hand Reconstruction Using Inverse Rendering (AAAI 2024 · similarity 0.88)
- BiTT: Bi-directional Texture Reconstruction of Interacting Two Hands from a Single Image (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.
