2020/03/02 by D.R.O. Morrison, Morrison, Douglas, Peter Corke +3 · 9 citations
Computer Science · Engineering · #FOS: Computer and information sciences #Human Pose and Action Recognition #Multimodal Machine Learning Applications #Robot Manipulation and Learning #Robotics (cs.RO)
paper · pdf · doi:10.48550/arxiv.2003.01314
openalex publication_date 2020/03/02 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We present the Evolved Grasping Analysis Dataset (EGAD), comprising over 2000\ngenerated objects aimed at training and evaluating robotic visual grasp\ndetection algorithms. The objects in EGAD are geometrically diverse, filling a\nspace ranging from simple to complex shapes and from easy to difficult to\ngrasp, compared to other datasets for robotic grasping, which may be limited in\nsize or contain only a small number of object classes. Additionally, we specify\na set of 49 diverse 3D-printable evaluation objects to encourage reproducible\ntesting of robotic grasping systems across a range of complexity and\ndifficulty. The dataset, code and videos can be found at\nhttps://dougsm.github.io/egad/\n