Automated robot grasping learns from skeleton-based geometry guidance
Skeleton-Guided Geometry-Aware Auto-Labeling for Robotic Grasping: From 4-DoF Learning to 6-DoF Execution
Robotics
Summary
Robotic grasping usually needs lots of human-labeled examples to work well, which is slow and expensive. The authors present a new way to automatically create labels for robot grasps using a skeleton-like shape representation of objects. They build a lightweight learning system that uses these labels for robot arms to pick up objects precisely in 3D space. Tests show their method works well on different datasets and in real-world scenes, even cluttered ones, often beating existing methods without extra tuning. This approach helps robots learn to grasp better without relying on manual labeling.
What this means in practice
- •For robotic automation engineers: Create precise grasp annotations automatically to train robots for picking objects reliably in both simple and cluttered 3D environments.
- •For warehouse automation operators: Use lightweight grasp learning models trained on auto-labeled data to improve robot picking success rates without extensive manual dataset creation.
Authors
Ali Sabzejou, Farbod Azimmohseni, Mehdi Tale Masouleh, Ahmad Kalhor
Abstract
Robotic grasping relies on large manually annotated datasets, which are costly to create. We introduce a straight-skeleton-based algorithm for generating two-finger grasp keypoints and identifying regions suitable for stable, feasible grasps. Our automated pipeline produces consistent grasp annotations without human intervention. We also propose the Skeleton-based Generative Grasp Convolutional Neural Network (SkelGG-CNN), a 4-DoF architecture that learns geometric grasp features from skeleton-based guidance. To extend auto-labeling to 6-DoF grasping, we develop a generalized quasi-3D skeletonization algorithm that operates directly on point clouds. Using automatically generated 4-DoF labels, SkelGG-CNN achieves accuracies of 95.42%, 96.55%, and 97.1% on Jacquard V1, Jacquard V2, and Cornell, respectively. On the GraspNet novel split, 6-DoF auto-labeling achieves an average precision of 18.33, comparable to state-of-the-art model-based approaches. In simulation, the 6-DoF method achieves success rates of 98.4% on Dex-Net and 89.1% on EGAD!. Real-world experiments with a Delta parallel robot show success rates of 97.4% on single household objects and 95.0% in cluttered scenes, outperforming state-of-the-art methods without domain adaptation. The 4-DoF pipeline adapts across environments and produces annotations approaching human-level precision, enabling efficient training of lightweight models on small datasets. The generalized 6-DoF method provides scalable, geometry-aware annotations for cluttered 3D scenes.