Multi-fingered Hand Grasping Dataset

该数据集是一个针对杂乱环境中多指手抓取的综合性数据集，包含了杂乱场景的点云数据、抓取质量以及可供性信息。此外，数据集还包含了手动标注的对象可供性标签，如握把抓取、环绕抓取、倾倒、按压、切割和扭转等。该数据集规模宏大，由1,521个家用物品模型生成了2,000个场景，其任务专注于在杂乱场景中的抓取研究。

This comprehensive dataset is designed for multifingered hand grasping research in cluttered environments. It encompasses point cloud data of cluttered scenes, grasp quality metrics, and affordance information. Additionally, the dataset contains manually annotated object affordance labels, including power grasp, wrap grasp, pour, press, cut, and twist. It has a large scale, with 2,000 cluttered scenes generated from 1,521 household object models, and its tasks focus on grasping studies in cluttered environments.