Research map / 3D vision / Human pose estimation
Reconstructing Interacting Hands with Interaction Prior from Monocular Images
Binghui Zuo, Zimeng Zhao, Wenqian Sun, Wei Xie, Zhou Xue, Yangang Wang
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 |
|---|---|---|---|
| build the interaction prior from multimodal data (marker-based Two-hand 500K, marker-less, hand-object) instead of only the paired benchmark data (ablation) | interaction prior · Interacting hands benchmark | MPJPE (mm): 8.77 -> 8.34 | Improved |
| add interaction prior instead of MLP regression (ablation) | interacting hands reconstruction · Interacting hands benchmark | MPJPE (mm): 11.07 -> 8.34 | Improved |
| add local feature extraction module (vs global only) (ablation) | interacting hands reconstruction · Interacting hands benchmark | MPJPE (mm): 10.19 -> 8.34 | Improved |
| add instance saliency maps (ablation) | feature extraction · Interacting hands benchmark | MPJPE (mm): 9.06 -> 8.34 | Improved |
| add interaction adjacency heatmap (IAH) (ablation) | feature extraction · Interacting hands benchmark | MPJPE (mm): 9.87 -> 8.34 | Improved |
| use IAH instead of all-in-one or joint-wise heatmaps (ablation) | feature extraction · Interacting hands benchmark | MPJPE (mm): 9.63 / 8.81 -> 8.34 | Improved |
| use Laplacian kernel instead of Gaussian in IAH (ablation) | IAH · Interacting hands benchmark | MPJPE (mm): 9.22 -> 8.34 | Improved |
| use ViT fusion instead of ResNet50 or HRNet32 (ablation) | feature fusion · Interacting hands benchmark | MPJPE (mm): 9.22 / 9.16 -> 8.34 | Improved |
| use MANO parameters instead of 3D joints or 3D vertices for the prior (ablation) | interaction prior · Interacting hands benchmark | MPJPE / MPVPE (mm): 9.74 (joints) -> 8.34; 10.52 (vertices) -> 8.51 MPVPE | Improved |
Closest papers by meaning
- MagicHOI: Leveraging 3D Priors for Accurate Hand-object Reconstruction from Short Monocular Video Clips (ICCV 2025 · similarity 0.89)
- Reconstructing Hand-Object Interactions in the Wild (ICCV 2021 · similarity 0.89)
- BIGS: Bimanual Category-agnostic Interaction Reconstruction from Monocular Videos via 3D Gaussian Splatting (CVPR 2025 · similarity 0.89)
- EasyHOI: Unleashing the Power of Large Models for Reconstructing Hand-Object Interactions in the Wild (CVPR 2025 · similarity 0.89)
- HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video (CVPR 2024 · similarity 0.88)
- MEgoHand: Multimodal Egocentric Hand-Object Interaction Motion Generation (NeurIPS 2025 · 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.
