... Recognition in context; Superpixel stuff segmentation; 330K images (>200K ... 91 stuff categories; 5 captions per image; 250,000 people with keypoints ...
確定! 回上一頁