Research map / 3D vision / 3D generation
Grasp’D: Differentiable Contact-Rich Grasp Synthesis for Multi-Fingered Hands
Dylan Turpin, Liquan Wang, Eric Heiden, Yun-Chun Chen, Miles Macklin, Stavros Tsogkas, Sven Dickinson, Animesh Garg
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 problem relaxation with extra force variables (ablation) | Grasp'D · YCB MANO grasps (21 objects) | contact area: 6.1 -> 42.6 | Improved |
| add problem relaxation with extra force variables (ablation) | Grasp'D · YCB MANO grasps (21 objects) | simulation displacement: 3.82 -> 0.41 | Improved |
| add problem relaxation with extra force variables (ablation) | Grasp'D · YCB MANO grasps (21 objects) | Ferrari-Canny epsilon: 0.52 -> 2.38 | Improved |
| add coarse-to-fine surface smoothing (ablation) | Grasp'D · YCB MANO grasps (21 objects) | simulation displacement: 0.55 -> 0.41 | Improved |
| add coarse-to-fine surface smoothing (ablation) | Grasp'D · YCB MANO grasps (21 objects) | contact area: 43.2 -> 42.6 | No clear effect |
| replace analytic GraspIt! synthesis with differentiable simulation | Grasp'D (top2) · ShapeNet MANO grasps | contact area: 9.4 -> 43.0 | Improved |
| replace analytic GraspIt! synthesis with differentiable simulation | Grasp'D (top2) · ShapeNet MANO grasps | simulation displacement: 1.95 -> 0.59 | Improved |
| replace analytic GraspIt! synthesis with differentiable simulation | Grasp'D (top2) · ShapeNet MANO grasps | intersection volume: 1.28 -> 5.70 | Hurt |
| replace analytic GraspIt! synthesis with differentiable simulation | Grasp'D (top2) · ShapeNet MANO grasps | Ferrari-Canny epsilon: 4.70 -> 5.01 | Improved |
| synthesize from RGB-D reconstructions instead of ground-truth meshes | Grasp'D · YCB MANO grasps | simulation displacement: 0.41 -> 5.68 | Hurt |
| fine-tune Grasping Field on Grasp'D data instead of GraspIt! data | Grasping Field · unseen YCB objects | contact area: 15.78 -> 21.00 | Improved |
| fine-tune Grasping Field on Grasp'D data instead of GraspIt! data | Grasping Field · unseen YCB objects | intersection volume: 9.44 -> 11.23 | Hurt |
Closest papers by meaning
- D-Grasp: Physically Plausible Dynamic Grasp Synthesis for Hand-Object Interactions (CVPR 2022 · similarity 0.93)
- Force-Aware 3D Contact Modeling for Stable Grasp Generation (AAAI 2026 · similarity 0.91)
- Contact Map Transfer with Conditional Diffusion Model for Generalizable Dexterous Grasp Generation (NeurIPS 2025 · similarity 0.90)
- CoToGrasp: Contact-Topology-Conditioned Dexterous Grasp Synthesis via Canonical Workspace Learning (ECCV 2026 · similarity 0.90)
- ContactOpt: Optimizing Contact To Improve Grasps (CVPR 2021 · similarity 0.90)
- 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.
