Research map / Image generation / Text-to-image diffusion
G-DexGrasp: Generalizable Dexterous Grasping Synthesis Via Part-Aware Prior Retrieval and Prior-Assisted Generation
Juntao Jian, Xiuping Liu, Zixuan Chen, Manyi Li, Jian Liu, Ruizhen Hu
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 |
|---|---|---|---|
| replace OBB part-size descriptor with BPS part encoding for prior retrieval (ablation) | G-DexGrasp · Unseen-category test set | Penetration depth: 0.29 -> 0.38 | No clear effect |
| apply pipeline to open-set Objaverse objects segmented by PartField | G-DexGrasp · Objaverse-tiny (500 objects) | Part accuracy and quality: Solid intersection volume 2.94 -> 5.93; part accuracy 71.6 -> 70.0 | Mixed |
| reduce intrinsic prior set to 30 per affordance type (about 1/100) (ablation) | G-DexGrasp · Unseen-category test set | Simulation displacement: 4.27 -> 5.91; intersection volume 2.94 -> 2.35 | Mixed |
| train with only 1/10 of extrinsic parameters (ablation) | G-DexGrasp · Unseen-category test set | Solid intersection volume: 2.94 -> 4.25 | Hurt |
| remove retrieved prior guidance entirely (ablation) | G-DexGrasp · Unseen-category test set | grasp quality | Hurt |
| add refinement optimization after generation (ablation) | G-DexGrasp · Unseen-category test set | penetration and stability (qualitative) | Improved |
| use G-DexGrasp instead of GraspTTA | Grasp synthesis · AffordPose test set | Rotation std R_std: 65.51 -> 78.57 | Improved |
| use G-DexGrasp instead of GraspTTA on joint angle diversity | Grasp synthesis · AffordPose test set | Joint angle std J_std: 17.43 -> 13.18 | Hurt |
| add point and press affordance types (7 instead of 5) (ablation) | GPT-4o affordance parsing · Unseen data | affordance prediction accuracy: 96.99 -> 96.66 | No clear effect |
Closest papers by meaning
- RAGrasp: A Retrieval-Augmented Framework with Diversity-Aware Modeling for Dexterous Grasp Generation (ECCV 2026 · similarity 0.87)
- UniDexGrasp: Universal Robotic Dexterous Grasping via Learning Diverse Proposal Generation and Goal-Conditioned Policy (CVPR 2023 · similarity 0.87)
- DexGrasp Anything: Towards Universal Robotic Dexterous Grasping with Physics Awareness (CVPR 2025 · similarity 0.87)
- CADGrasp: Learning Contact and Collision Aware General Dexterous Grasping in Cluttered Scenes (NeurIPS 2025 · similarity 0.85)
- Contact Map Transfer with Conditional Diffusion Model for Generalizable Dexterous Grasp Generation (NeurIPS 2025 · similarity 0.85)
- Dexterous Grasp Transformer (CVPR 2024 · similarity 0.85)
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.
