What’s Out There
← Back to swiping
Images & videoFreeAcademic

ImageNet

Stanford University and Princeton University (ImageNet project)

The 14-million-photo set whose 2012 contest kicked off the deep-learning boom.

Photos gathered from the web and labeled by crowdworkers against the WordNet noun hierarchy. The best-known slice, ILSVRC-2012, has about 1.28 million training images in 1,000 classes and was the benchmark on which neural networks first beat other methods decisively. Its 'person' categories were later found to include offensive labels and were pruned.

Access

Cost
FreeFree after creating an account and agreeing to the terms of access.
License
Open, non-commercial onlyTerms of access limit use to non-commercial research and education; image copyrights stay with owners
Sign-up
Free account
How you get it
Bulk downloadJPEG, XML (boxes)

What's in it

One row is
One image labeled with a WordNet synset (and bounding boxes for a subset)
Covers
Web images, skewed toward North America and Europe
Years
Collected 2007-2010; ILSVRC contests 2010-2017
Updated
One-time release
Size
About 14 million images in about 21,000 categories; ILSVRC-2012 has about 1.28 million training images
imagesynset IDclass labelbounding boxsplit

Quality

4 of 5

Historically pivotal and still a standard pretraining and evaluation set, with well-documented label noise and representational problems.

  • Person categories contained offensive labels and were removed in 2019
  • Several percent of validation labels are wrong or ambiguous
  • Images include identifiable people photographed without consent

Connects to

  • COCO (Common Objects in Context): Adds object locations and captions for scenes, where ImageNet has one label per image
  • Open Images V7: A larger, openly licensed alternative with boxes and masks
  • LAION-5B: Web-scale image-text pairs that replaced labeled sets like ImageNet for pretraining

Questions it could answer

  • Which classes are most often confused, and are they genuinely ambiguous?
  • How geographically diverse are the photos in household-object categories?

Similar data sets

  • FineVision

    17.3 million images and 24.3 million samples merged from many sources to train open vision-language models.

    Images & videoHugging Face (M4 team)Free

  • COCO (Common Objects in Context)

    330,000 everyday photos with every person, dog, and fork outlined by hand.

    Images & videoCOCO Consortium (Microsoft and academic partners)Free

  • Open Images V7

    About 9 million Creative Commons photos with 16 million boxes and 61 million labels.

    Images & videoGoogleFree

  • MedMNIST+: 18 standardized biomedical image datasets

    Eighteen medical imaging classification tasks shrunk to MNIST-size, in several resolutions.

    Medicine & clinicalYang, Shi, Wei, et al. (Shanghai Jiao Tong University)Free

  • PD12M

    12.4 million public-domain and CC0 image-caption pairs for training image models.

    Images & videoSpawningFree

  • Kinetics-700

    650,000 ten-second YouTube clips of people doing 700 things, from yoga to changing a tire.

    Images & videoGoogle DeepMind (hosted by the Common Visual Data Foundation)Free