The library uses python, which must be installed. Python 2.7, 3.4, 3.5 and 3.6 are supported.

The following packages must be installed:

  • numpy
  • scipy
  • scikit-image (pip install -U scikit-image)
  • six (pip install -U six)
  • OpenCV (i.e. cv2 must be available in python). The library is mainly tested in OpenCV 2, but seems to also work in OpenCV 3.

When executing the installer, these packages will be automatically installed/upgraded where needed. This is not the case for OpenCV, which must be installed manually.

Once the packages are available, imgaug can be installed using the following command:

pip install git+

This installs the latest version directly from github. If any error pops up, try adding sudo in front (i.e. sudo pip install git+

Alternatively, you can install the latest version which was added to pypi via pip install imgaug. That version can sometimes be behind the version on github.

In rare cases, one might prefer to install from the locally cloned repository. This is possible using python sdist && sudo pip install dist/imgaug-VERSION.tar.gz, where VERSION must be replaced by the current version of the library (e.g. imgaug-0.2.6.tar.gz). The current version can be derived from’s content.

To deinstall the library use pip uninstall imgaug.