Not logged in. Login | Signup

ImageNet Large Scale Visual Recognition Challenge 2015 (ILSVRC2015)

Back to Main page  

Citation

When using the DET or CLS-LOC dataset, please cite:
    Olga Russakovsky*, Jia Deng*, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, Alexander C. Berg and Li Fei-Fei. (* = equal contribution) ImageNet Large Scale Visual Recognition Challenge. arXiv:1409.0575, 2014. paper | bibtex
When using the Places2 dataset for the taster scene classification challenge, please cite:
    Bolei Zhou, Aditya Khosla, Agata Lapedriza, Antonio Torralba and Aude Oliva. Places2: A Large-scale Database for Scene Understanding. Arxiv, 2015. (coming soon)

Main competitions

Development kit

Please be sure to consult the included readme.txt file for competition details. Additionally, the development kit includes

  • Overview and statistics of the data.
  • Meta data for the competition categories.
  • Matlab routines for evaluating submissions.

Object classification/localization


This dataset is unchanged from ILSVRC2012. There are a total of 1,281,167 images for training. The number of images for each synset (category) ranges from 732 to 1300. There are 50,000 validation images, with 50 images per synset. There are 100,000 test images. All images are in JPEG format.

Object detection


    DET dataset. 49GB. MD5: 676c745e4329b0592cd855ec0fbbae94

    The training and validation dataset is unchanged from ILSVRC2014. There are a total of 456567 images for training. The number of positive images for each synset (category) ranges from 461 to 67513. The number of negative images ranges from 42945 to 70626 per synset. There are 20121 validation images. All images are in JPEG format.

    DET test dataset. 6GB. MD5: 51e04a189e97dece83a025d69c3888e8

    The test dataset is refreshed by adding 11142 new images. The file contains 51294 (40152 + 11142) images and the test.txt file.

    DET test dataset (new). 914MB. MD5: 9d7685ed0f26281ac87e31d82f036be1

    The file only contains 11142 new images and the test.txt file.

Taster competitions

Object detection from video (VID)


    Please download from here and read carefully the term of use for the VID task.

Scene classification


    Please refer to the places2 website for the download of the latest version of Places2 dataset. The previous version of Places2, termed as Places401, is deprecated and not supported anymore.
    Places2 Development kit

    Please be sure to consult the included README file for competition details. Additionally, the development kit includes

    • Overview and statistics of the data.
    • Meta data for the competition categories.
    • Matlab routines for evaluating submissions.

    High-resolution images

    Train images. 479GB. MD5: 55cf2c4a727b0ed568290456686d0bd0

    Validation images. 1.2GB. MD5: 1c925f2d9d8e1fe0498981fe2bd66687

    Test images. 22GB. MD5: 24af6f4cb5da3c55267f74be84872b5c

    The images in the above archives have been resized to have a minimum dimension of 512 while preserving the aspect ratio of the image. Original images that had a dimension smaller than 512 have been left unchanged.

    Small images (256 * 256)

    Train images. 109GB. MD5: b93e273dfb37e2243add2547ed2f38d3

    Validation images. 275MB. MD5: 3258f32ebe3038873e5a96bbf1253649

    Test images. 5.1GB. MD5: 24b9043f84a0418854e102ea6cb60b3b

    The images in the above archives have been resized to 256 * 256 regardless of the original aspect ratio.

Terms of use: by downloading the image data from the above URLs, you agree to the following terms:

  1. You will use the data only for non-commercial research and educational purposes.
  2. You will NOT distribute the above URL(s).
  3. Stanford University and Princeton University and UNC Chapel Hill and MIT make no representations or warranties regarding the data, including but not limited to warranties of non-infringement or fitness for a particular purpose.
  4. You accept full responsibility for your use of the data and shall defend and indemnify Stanford University and Princeton University and UNC Chapel Hill and MIT, including their employees, officers and agents, against any and all claims arising from your use of the data, including but not limited to your use of any copies of copyrighted images that you may create from the data.