General grasp-pose detector from the GraspNet team (SJTU): dense, temporally-smooth 7-DoF grasp poses on unseen objects from a single point cloud — reported ~93% bin-picking success, distributed as a license-gated SDK, free for research.
Input
point cloud
Output
6-DOF grasp pose
General grasp-pose detector from the GraspNet team (SJTU): dense, temporally-smooth 7-DoF grasp poses on unseen objects from a single point cloud — reported ~93% bin-picking success, distributed as a license-gated SDK, free for research.
Save it, compare it, request a quote or record that you have used it — with an account.
AnyGrasp, from the Machine Vision and Intelligence Group at Shanghai Jiao Tong University (the team behind the GraspNet-1Billion benchmark), predicts dense 7-DoF parallel-jaw grasp poses over full scenes from a single-view point cloud, with grasp quality scores and temporal smoothness for tracking grasps on moving objects. The T-RO 2023 paper reports human-comparable grasping with ~93% success in cluttered bin-picking of unseen objects. It is the most widely deployed academic grasp detector in research stacks (often chained after Grounding DINO + SAM for language-conditioned picking). Distribution is a compiled SDK with machine-locked license files — free for academic/research use, commercial licensing by contacting the authors.
This product is part of the Birdwave Atlas community.
Contribute evaluations, share integrations and help improve compatibility data.
Discussions, evaluations, integrations and updates