An algorithm using Ant Colony Optimization with polygons and ray casting for classification.

The easiest way to install the project is by using Conda. Conda is a Python package manager and environment manager that makes it easy to set up and install Python environments. Follow the instructions on Condas website to get started.

After installing Conda, cd into the project directory and install the project environment with:

$ conda env create -f environment.yml python=3.5

This will install a full conda environment named acoc. Activate the enviroment by running:

$ source activate acoc

The project looks for a configuration file called config.py in the root folder. This file is ignored by the VCS so that you can make changes without affecting the repository. To get started with a config file simply copy config_template.py and rename it to config.py. Then you can freely change the contents of the config file without affecting the VCS.

The algorithm is optimized for Nvidia GPUs and depends on CUDA. To run the project without CUDA, set 'gpu': false in config.py.

A complete classification example can be found in demo.py in the root folder of the repository. The demo use the classifier configuration specified in config.py.