> ML_DATASET // MSCOCO-2017_v1.0
MS COCO 2017 (Common Objects in Context)
COCO Consortium / Microsoft Research · Computer Vision & Scene Understanding · 118,287 training images, 80 object categories, 5,000 validation images
Computer Vision & Scene UnderstandingCC-BY-4.0 (Annotations) / Flickr Terms (Images)118,287 training images, 80 object categories, 5,000 validation imagesopen
Dataset Profile & Characteristics
Label Type:Bounding boxes, segmentation masks, keypoints, captions
Languages:en
License Tier:permissive-open-source
Modalities:image
Intended Use
- Object detection, instance segmentation, panoptic segmentation
- Keypoint estimation and image captioning
Prohibited / Discouraged Use
- Targeting private individuals
Bias, Leakage & Privacy Risk Analysis
Privacy / Sensitive Data Risks:
Flickr photos depicting people in public settings.
Known Bias:
Photographer perspective bias; indoor and outdoor consumer scene composition.
Known Benchmark Leakage:
Clean train/val split maintained by official evaluation server.