Research map / 3D vision / 3D generation
Hand-Object Contact Consistency Reasoning for Human Grasps Generation
Hanwen Jiang, Shaowei Liu, Jiashun Wang, Xiaolong 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 |
|---|---|---|---|
| add hand-centric loss L_H (ablation) | GraspCVAE · Obman | grasp displacement mean (cm): 3.51 -> 1.72 | Improved |
| add hand-centric loss L_H (ablation) | GraspCVAE · Obman | penetration depth (cm): 0.40 -> 0.48 | Hurt |
| add object-centric loss L_O on top of L_H (ablation) | GraspCVAE · Obman | grasp displacement mean (cm): 1.72 -> 1.63 | Improved |
| find contact regions dynamically with L_H instead of forcing ground-truth contact L_H(gt) (ablation) | GraspCVAE · Obman | grasp displacement mean (cm): 1.97 -> 1.72 | Improved |
| use normalized contact map in L_O instead of raw distance (ablation) | GraspCVAE · Obman | grasp displacement mean (cm): 1.77 -> 1.63 | Improved |
| add hand input to ContactNet with global features only (ablation) | ContactNet · contact map error | error: 0.161 -> 0.148 | Improved |
| add object local features to ContactNet with hand input (ablation) | ContactNet · contact map error | error: 0.148 -> 0.090 | Improved |
| add test-time adaptation (offline TTA) (ablation) | GraspCVAE · HO-3D | grasp displacement mean (cm): 4.98 -> 3.80 | Improved |
| make TTA online across video sequence instead of offline (ablation) | GraspCVAE · HO-3D | grasp displacement mean (cm): 3.80 -> 3.21 | Improved |
| train ContactNet with injected hand noise (TTA-noise) instead of clean data (ablation) | GraspCVAE TTA · HO-3D | grasp displacement mean (cm): 3.80 -> 4.22 | Hurt |
| optimize MANO parameters directly (TTA-optm) instead of finetuning the decoder (ablation) | GraspCVAE TTA · HO-3D | grasp displacement mean (cm): 3.80 -> 4.14 | Hurt |
| add online TTA (ablation) | GraspCVAE · FPHA | grasp displacement mean (cm): 2.55 -> 2.77 | No clear effect |
Closest papers by meaning
- ContactGen: Generative Contact Modeling for Grasp Generation (ICCV 2023 · similarity 0.91)
- Force-Aware 3D Contact Modeling for Stable Grasp Generation (AAAI 2026 · similarity 0.90)
- Contact Map Transfer with Conditional Diffusion Model for Generalizable Dexterous Grasp Generation (NeurIPS 2025 · similarity 0.89)
- GraspXL: Generating Grasping Motions for Diverse Objects at Scale (ECCV 2024 · similarity 0.89)
- RAGG: Retrieval-Augmented Grasp Generation Model (AAAI 2025 · similarity 0.88)
- LatentHOI: On the Generalizable Hand Object Motion Generation with Latent Hand Diffusion. (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.
