RED

홈페이지 · 논문 4편

The Real Embodied Dataset (RED) is a computer vision large-scale dataset for grasping in cluttered scenes. It contains complete segmentation masks for partially occluded objects, with their order of occlusion. Source: https://arxiv.org/pdf/2004.13358.pdf Image Source: https://arxiv.org/pdf/2004.13358.pdf