This project sets out to implement a generic image classification library that can take a training dataset of images and learn to identify if they belong to a certain category or not. It comes with an API as well as a Web interface, so that the user can see similarities and other metrics after training a model.
Requirements:
- numpy
- scipy
- sklearn
To install the library, one must have a running Ubuntu or Debian-based environment and run install_dependencies.sh.
You can also install it on Windows, providing you install the above libraries before running install_library.cmd. I suggest not using virtualenvs when installing on windows.
The install script works if you start by creating a virtual environment, by doing the following:
- sudo apt-get install virtualenv
- virtualenv --system-site-packages virtualenv
After that, run install_library.sh and you will have the imclas package built in the library.