Research map / 3D vision / 3D generation
Learning Joint Reconstruction of Hands and Manipulated Objects
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Balance attraction and repulsion contact losses (ablation) | Joint hand-object reconstruction · ObMan | simulation displacement (mm): 31.3 -> 30.9 | Helped |
| Balance attraction and repulsion contact losses (ablation) | Joint hand-object reconstruction · FHB_C | maximum penetration (mm): 18.7 ±0.6 -> 12.1 ± 0.7 | Helped |
| Use attraction loss only (ablation) | Joint hand-object reconstruction · ObMan | maximum penetration (mm): 9.5 -> 11.8 | Hurt |
| Use attraction loss only (ablation) | Joint hand-object reconstruction · ObMan | simulation displacement (mm): 31.3 -> 26.8 | Helped |
| Use repulsion loss only (ablation) | Joint hand-object reconstruction · ObMan | maximum penetration (mm): 9.5 -> 6.4 | Helped |
| Use repulsion loss only (ablation) | Joint hand-object reconstruction · ObMan | simulation displacement (mm): 31.3 -> 38.1 | Hurt |
| Train with hand-object occlusions rather than isolated hands (ablation) | Hand reconstruction · ObMan hand-object images | hand error: 14.1 -> 11.6 | Helped |
| Train with hand-object occlusions rather than isolated objects (ablation) | Object reconstruction · ObMan hand-object images | object error: 0.0722 -> 0.0302 | Helped |
| Add hand-shape regularization to joint supervision (ablation) | MANO hand branch · FHB | mean endpoint error (mm): 28.1 -> 26.5 | Helped |
| Add vertex supervision to joints and shape regularization (ablation) | MANO hand branch · ObMan | mean endpoint error (mm): 11.7 -> 11.6 | Helped |
| Use 45 rather than 30 pose PCA components (ablation) | MANO hand branch · ObMan | mean endpoint error (mm): 11.6 -> 11.2 | Helped |
| Predict objects in camera view rather than canonical view (ablation) | AtlasNet object branch · Object reconstruction | Chamfer loss (× 1000): 4.88 -> 4.88 | No clear effect |
Closest papers by meaning
- HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video (CVPR 2024 · similarity 0.91)
- Learning Explicit Contact for Implicit Reconstruction of Hand-Held Objects from Monocular Images (AAAI 2024 · similarity 0.90)
- ComPose: When to Trust Hands for Object Pose Tracking (NeurIPS 2026 · similarity 0.89)
- ManiVideo: Generating Hand-Object Manipulation Video with Dexterous and Generalizable Grasping (CVPR 2025 · similarity 0.89)
- G-HOP: Generative Hand-Object Prior for Interaction Reconstruction and Grasp Synthesis (CVPR 2024 · similarity 0.89)
- DiffGrasp: Whole-Body Grasping Synthesis Guided by Object Motion Using a Diffusion Model (AAAI 2025 · similarity 0.89)
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.