Tiger stripe database submissions

An open-call for tiger images. The EIA Tiger Stripes project aims to develop a tiger stripe identification AI tool to help identify individual tigers from their stripe pattern profiles.

Tiger stripe patterns are as unique as human fingerprints and in order to develop, train and test the technology we need thousands of images of individual tiger stripe patterns, sourced by EIA staff, other organisations and you! The ultimate ambition is to create a globally accessible AI tool and reference database to assist law enforcement and researchers to cross-reference tiger stripe patterns from skins seen for sale or seized in trade, and live tigers. Read more here.

 

  • e.g country, state/province, name of designated protected area, name of forest area, or zoo/rescue centre
  • If it’s a wild tiger, the official name e.g. T-22 and popular name e.g. Sultan
    If it’s a captive tiger, the popular or given name if there is one
  • Drop files here or

    Max. file size: 63 MB.

      To select multiple images hold control on PC or command on Mac
    • By ticking this box, EIA commits to use your image solely for the purposes of the partnership with the Alan Turing Institute to develop, train and test new AI technology aimed at identifying individual tigers from their stripe patterns. This project will involve creating bounding boxes around the image you submit. The project will be initially implemented by an ATI sponsored doctoral researcher for six months and may be followed with further examination under an ATI Data Study Group. EIA will not use your image for any other purpose without prior informed consent.
    • By ticking this box, EIA commits to use your images as described above at 1), and you agree that we may retain your image for the creation of a future database for cross-referencing future images of tiger skins seen for sale or seized in trade, which third parties may access but not download images from. EIA will not use your image for any other purpose without prior informed consent.

    Related