HANDAL: A Dataset of Real-World Manipulable Object Categories with Pose Annotations, Affordances, and Reconstructions
This dataset addresses the need for practical robotics research by providing manipulable object data for tasks like functional grasping, though it is incremental as it builds on existing dataset efforts.
The authors introduced HANDAL, a dataset of 308k annotated images from 2.2k videos covering 212 real-world manipulable objects in 17 categories, designed for robotics-ready category-level pose estimation and affordance prediction. They developed a streamlined annotation process using a single camera and semi-automated processing to produce high-quality 3D annotations without crowd-sourcing.
We present the HANDAL dataset for category-level object pose estimation and affordance prediction. Unlike previous datasets, ours is focused on robotics-ready manipulable objects that are of the proper size and shape for functional grasping by robot manipulators, such as pliers, utensils, and screwdrivers. Our annotation process is streamlined, requiring only a single off-the-shelf camera and semi-automated processing, allowing us to produce high-quality 3D annotations without crowd-sourcing. The dataset consists of 308k annotated image frames from 2.2k videos of 212 real-world objects in 17 categories. We focus on hardware and kitchen tool objects to facilitate research in practical scenarios in which a robot manipulator needs to interact with the environment beyond simple pushing or indiscriminate grasping. We outline the usefulness of our dataset for 6-DoF category-level pose+scale estimation and related tasks. We also provide 3D reconstructed meshes of all objects, and we outline some of the bottlenecks to be addressed for democratizing the collection of datasets like this one.