COCO-Stuff

Common Objects in COntext-stuff

ImagesVariousIntroduced 2018-01-01

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

Benchmarks

10-shot image generation/F.W. IU 10-shot image generation/Per-Class Accuracy 10-shot image generation/Pixel Accuracy 10-shot image generation/mIoU 10-shot image generation/Frame (fps)Semantic Segmentation/F.W. IU Semantic Segmentation/Per-Class Accuracy Semantic Segmentation/Pixel Accuracy Semantic Segmentation/mIoU Semantic Segmentation/Frame (fps)Sketch-to-Image Translation/FID Sketch-to-Image Translation/FID-C Zero-Shot Semantic Segmentation/Transductive Setting hIoU Zero-Shot Semantic Segmentation/Inductive Setting hIoU

Related Benchmarks

COCO-Stuff 128x128/Image Generation/FID COCO-Stuff 128x128/Image Generation/Inception Score COCO-Stuff 128x128/Image Generation/SceneFID COCO-Stuff 256x256/Image Generation/FID COCO-Stuff 256x256/Image Generation/Inception Score COCO-Stuff 256x256/Image Generation/LPIPS COCO-Stuff 64x64/Image Generation/FID COCO-Stuff 64x64/Image Generation/Inception Score COCO-Stuff Labels-to-Photos/1 Image, 2*2 Stitching/Accuracy COCO-Stuff Labels-to-Photos/1 Image, 2*2 Stitching/FID COCO-Stuff Labels-to-Photos/1 Image, 2*2 Stitching/LPIPS COCO-Stuff Labels-to-Photos/1 Image, 2*2 Stitching/mIoU COCO-Stuff Labels-to-Photos/Image Generation/Accuracy COCO-Stuff Labels-to-Photos/Image Generation/FID COCO-Stuff Labels-to-Photos/Image Generation/LPIPS COCO-Stuff Labels-to-Photos/Image Generation/mIoU COCO-Stuff Labels-to-Photos/Image-to-Image Translation/Accuracy COCO-Stuff Labels-to-Photos/Image-to-Image Translation/FID COCO-Stuff Labels-to-Photos/Image-to-Image Translation/LPIPS COCO-Stuff Labels-to-Photos/Image-to-Image Translation/mIoU COCO-Stuff full/10-shot image generation/Mean IoU (class)COCO-Stuff full/Semantic Segmentation/Mean IoU (class)COCO-Stuff test/10-shot image generation/mIoU COCO-Stuff test/Semantic Segmentation/mIoU COCO-Stuff val/10-shot image generation/mIoU COCO-Stuff val/Semantic Segmentation/mIoU COCO-Stuff-15/10-shot image generation/Pixel Accuracy COCO-Stuff-15/Semantic Segmentation/Pixel Accuracy COCO-Stuff-15/Unsupervised Semantic Segmentation/Pixel Accuracy COCO-Stuff-171/10-shot image generation/Pixel Accuracy COCO-Stuff-171/10-shot image generation/mIoU COCO-Stuff-171/Open Vocabulary Semantic Segmentation/HIoU COCO-Stuff-171/Open Vocabulary Semantic Segmentation/mIoU COCO-Stuff-171/Semantic Segmentation/Pixel Accuracy COCO-Stuff-171/Semantic Segmentation/mIoU COCO-Stuff-171/Unsupervised Semantic Segmentation/Pixel Accuracy COCO-Stuff-171/Unsupervised Semantic Segmentation/mIoU COCO-Stuff-27/10-shot image generation/Clustering [Accuracy]COCO-Stuff-27/10-shot image generation/Clustering [mIoU]COCO-Stuff-27/10-shot image generation/Linear Classifier [Accuracy]COCO-Stuff-27/10-shot image generation/Linear Classifier [mIoU]COCO-Stuff-27/10-shot image generation/Pixel Accuracy COCO-Stuff-27/10-shot image generation/mIoU COCO-Stuff-27/10-shot image generation/pixel accuracy COCO-Stuff-27/Semantic Segmentation/Clustering [Accuracy]COCO-Stuff-27/Semantic Segmentation/Clustering [mIoU]COCO-Stuff-27/Semantic Segmentation/Linear Classifier [Accuracy]COCO-Stuff-27/Semantic Segmentation/Linear Classifier [mIoU]COCO-Stuff-27/Semantic Segmentation/Pixel Accuracy COCO-Stuff-27/Semantic Segmentation/mIoU COCO-Stuff-27/Semantic Segmentation/pixel accuracy COCO-Stuff-27/Unsupervised Semantic Segmentation/Clustering [Accuracy]COCO-Stuff-27/Unsupervised Semantic Segmentation/Clustering [mIoU]COCO-Stuff-27/Unsupervised Semantic Segmentation/Linear Classifier [Accuracy]COCO-Stuff-27/Unsupervised Semantic Segmentation/Linear Classifier [mIoU]COCO-Stuff-27/Unsupervised Semantic Segmentation/mIoU COCO-Stuff-27/Unsupervised Semantic Segmentation/pixel accuracy COCO-Stuff-3/10-shot image generation/Pixel Accuracy COCO-Stuff-3/Semantic Segmentation/Pixel Accuracy COCO-Stuff-3/Unsupervised Semantic Segmentation/Pixel Accuracy COCO-Stuff-81/10-shot image generation/Pixel Accuracy COCO-Stuff-81/10-shot image generation/mIoU COCO-Stuff-81/Semantic Segmentation/Pixel Accuracy COCO-Stuff-81/Semantic Segmentation/mIoU COCO-Stuff-81/Unsupervised Semantic Segmentation/Pixel Accuracy COCO-Stuff-81/Unsupervised Semantic Segmentation/mIoU