VMRD

홈페이지 · 논문 6편

VMRD is a multi-object grasp dataset. It has been collected and labeled using hundreds of objects coming from 31 categories. There are totally 5,185 images including 17,688 object instances and 51,530 manipulation relationships.