Common Objects in COntext-stuff
The Common Objects in COntext-stuff (COCO-stuff) dataset is a dataset for scene understanding tasks like semantic segmentation, object detection and image captioning. It is constructed by annotating the original COCO dataset, which originally annotated things while neglecting stuff annotations. There are 164k images in COCO-stuff dataset that span over 172 categories including 80 things, 91 stuff, and 1 unlabeled class.
Source: Image Colorization: A Survey and Dataset
Image Source: https://github.com/nightrome/cocostuff
Variants: COCO-Stuff-81, COCO-Stuff 256x256, COCO-Stuff-171, COCO-Stuff-27, COCO-Stuff full, COCO-Stuff-3, COCO-Stuff-15, COCO-Stuff test, COCO-Stuff Labels-to-Photos, COCO-Stuff 64x64, COCO-Stuff 128x128, COCO-Stuff
This dataset is used in 3 benchmarks:
Recent papers with results on this dataset: