DexGraspNet 2.0: Learning Generative Dexterous Grasping in Large-scale Synthetic Cluttered Scenes

CoRL 2024


Jialiang Zhang1,2,*    Haoran Liu1,2,*    Danshi Li2,*    Xinqiang Yu2,*    Haoran Geng1,2,3    Yufei Ding1,2    Jiayi Chen1,2    He Wang1,2,4,†

1CFCS, School of Computer Science, Peking University    2Galbot    3UC Berkeley    4Beijing Academy of Artificial Intelligence   

* equal contributions   corresponding author  


input

Simulation set : DexGraspNet 2.0 contains 427M grasps (4 random grasps are visualized in each scene here for clarity).
Real-world Execution : first row are model-generated grasps conditioned on real-world single-view depth point clouds, second row are top-ranked grasps, and third row are real-world executions.


Abstract


Grasping in cluttered scenes remains highly challenging for dexterous hands due to the scarcity of data. To address this problem, we present a large-scale synthetic benchmark, encompassing 1319 objects, 8270 scenes, and 427 million grasps. Beyond benchmarking, we also propose a novel two-stage grasping method that learns efficiently from data by using a diffusion model that conditions on local geometry. Our proposed generative method outperforms all baselines in simulation experiments. Furthermore, with the aid of test-time-depth restoration, our method demonstrates zero-shot sim-to-real transfer, attaining 90.7% real-world dexterous grasping success rate in cluttered scenes.


Video




set


Some synthetic training scenes and grasping pose from DexGraspNet 2.0.



Qualitative results


Dexterous grasping generation and real-world execution in a cluttered scene.



Citation



  @inproceedings{zhangdexgraspnet,
    title={DexGraspNet 2.0: Learning Generative Dexterous Grasping in Large-scale Synthetic Cluttered Scenes},
    author={Zhang, Jialiang and Liu, Haoran and Li, Danshi and Yu, XinQiang and Geng, Haoran and Ding, Yufei and Chen, Jiayi and Wang, He},
    booktitle={8th Annual Conference on Robot Learning}
  }

License


This work and the set are licensed under CC BY-NC 4.0.


Contact


If you have any questions, please feel free to contact Jialiang Zhang at zhangjialiang@stu.pku.edu.cn, Haoran Liu at lhrrhl0419@stu.pku.edu.cn, Danshi Li at danshi.li.academia@gmail.com, Xinqiang Yu at yuxinqiang@galbot.com, and He Wang at hewang@pku.edu.cn.