Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics

Short summary

Dex-Net 2.0 uses a deep learning model trained on 6.7 million synthetic point clouds to plan robotic grasps with 93% success on objects with adversarial geometry, outperforming point cloud registration methods.

AI-generated from the title and abstract; the full text is not read.