Research map / 3D vision / 3D generation
Force-Aware 3D Contact Modeling for Stable Grasp Generation
Zhuo Chen, Zhongqun Zhang, Yihua Cheng, Aleš Leonardis, Hyung Jin Chang
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 force-aware contact modeling and keypoint optimization vs GrabNet | force-aware grasp pipeline · GRAB | Simulation displacement (cm): 0.90 -> 0.68 | Improved |
| add force-aware pipeline vs best baseline distinction rate | force-aware grasp pipeline · GRAB | Distinction rate (%): 14.2 -> 17.5 | Improved |
| add force-aware pipeline on unseen HO3D objects vs ContactGen | force-aware grasp pipeline · HO3Dv2 (out-of-domain) | Distinction rate (%): 9.0 -> 17.0 | Improved |
| add force-aware pipeline on unseen HO3D objects vs GrabNet | force-aware grasp pipeline · HO3Dv2 (out-of-domain) | Simulation displacement (cm): 1.86 -> 1.50 | Improved |
| add force-aware pipeline vs GrabNet on penetration | force-aware grasp pipeline · GRAB | Penetration volume (cm^3): 5.13 vs 8.50 baseline; 5.13 vs 3.03 FastGrasp | Hurt |
| add force prediction branch in training (force map used in training only) (ablation) | force-aware grasp pipeline · GRAB | Simulation displacement (cm): 2.26 -> 2.02 | Improved |
| add force values as optimization weights (ablation) | force-aware grasp pipeline · GRAB | Simulation displacement (cm): 2.02 -> 1.60 | Improved |
| add keypoint-guided optimization (with average force) (ablation) | force-aware grasp pipeline · GRAB | Simulation displacement (cm): 2.02 -> 0.75 | Improved |
| add keypoints with predicted force values instead of average force (ablation) | force-aware grasp pipeline · GRAB | Distinction rate (%): 15.0 -> 17.5 | Improved |
| add predicted force values instead of average (penetration) (ablation) | force-aware grasp pipeline · GRAB | Penetration volume (cm^3): 6.54 -> 5.13 | Improved |
| use 3 keypoints instead of 0 (ablation) | force-aware grasp pipeline · GRAB | Simulation displacement (cm): 2.02 -> 0.68 | Improved |
| increase keypoints from 3 to 16 (ablation) | force-aware grasp pipeline · GRAB | Distinction rate (%): 17.5 -> 5.8 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ContactGen — ContactGen: Generative Contact Modeling for Grasp Generation (ICCV 2023) · better on 6 of 6 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
- Grasp’D: Differentiable Contact-Rich Grasp Synthesis for Multi-Fingered Hands (ECCV 2022 · similarity 0.91)
- D-Grasp: Physically Plausible Dynamic Grasp Synthesis for Hand-Object Interactions (CVPR 2022 · similarity 0.90)
- Contact Map Transfer with Conditional Diffusion Model for Generalizable Dexterous Grasp Generation (NeurIPS 2025 · similarity 0.90)
- RAGG: Retrieval-Augmented Grasp Generation Model (AAAI 2025 · similarity 0.87)
- VolCo: Volumetric Contact for High-Fidelity Human Grasp Generation (NeurIPS 2026 · similarity 0.87)
- CADGrasp: Learning Contact and Collision Aware General Dexterous Grasping in Cluttered Scenes (NeurIPS 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.
