{"work":{"id":"1adace8d-6bd2-4ecc-9769-afcd232a02dc","openalex_id":null,"doi":null,"arxiv_id":"1703.09312","raw_key":null,"title":"Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics","authors":null,"authors_text":null,"year":2017,"venue":"cs.RO","abstract":"To reduce data collection time for deep learning of robust robotic grasp plans, we explore training from a synthetic dataset of 6.7 million point clouds, grasps, and analytic grasp metrics generated from thousands of 3D models from Dex-Net 1.0 in randomized poses on a table. We use the resulting dataset, Dex-Net 2.0, to train a Grasp Quality Convolutional Neural Network (GQ-CNN) model that rapidly predicts the probability of success of grasps from depth images, where grasps are specified as the planar position, angle, and depth of a gripper relative to an RGB-D sensor. Experiments with over 1,000 trials on an ABB YuMi comparing grasp planning methods on singulated objects suggest that a GQ-CNN trained with only synthetic data from Dex-Net 2.0 can be used to plan grasps in 0.8sec with a success rate of 93% on eight known objects with adversarial geometry and is 3x faster than registering point clouds to a precomputed dataset of objects and indexing grasps. The Dex-Net 2.0 grasp planner also has the highest success rate on a dataset of 10 novel rigid objects and achieves 99% precision (one false positive out of 69 grasps classified as robust) on a dataset of 40 novel household objects, some of which are articulated or deformable. Code, datasets, videos, and supplementary material are available at http://berkeleyautomation.github.io/dex-net .","external_url":"https://arxiv.org/abs/1703.09312","cited_by_count":null,"metadata_source":"pith","metadata_fetched_at":"2026-07-04T06:29:37.592365+00:00","pith_arxiv_id":"1703.09312","created_at":"2026-05-17T20:55:52.199442+00:00","updated_at":"2026-07-04T06:29:37.592365+00:00","title_quality_ok":true,"display_title":"Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics","render_title":"Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics"},"hub":{"state":{"tier_text":"hub","tier":"hub","tier_reason":"10+ Pith inbound or 1,000+ external citations","pith_inbound_count":12,"external_cited_by_count":null},"tier":"hub","role_counts":[{"context_role":"background","n":1}],"polarity_counts":[{"context_polarity":"background","n":1}],"runs":{},"summary":{},"graph":{},"authors":[]}}