HANDGR is an Honours in Computer Science project conducted by Anna Borysova, Shaheel Kooverjee and Erin Versfeld. It makes use of three different gesture recognition devices, namely the Leap, the Myo and the Kinect, to gather data from skilled signers and to train various machine learning techniques on this data.
Each of the devices has it's own development environment in which the algorithms are trained. These are detailed below.
By cloning this repo and running gather_data.py one is able to collect data from all three devices simultaneously. This requires both Python2.7 and Python3 to be installed on your computer.
Have a bug or an issue with this project? Open a new issue here on GitHub or using the contact details available on our website.
HANDGR was created by and is maintained by Anna Borysova, Shaheel Kooverjee and Erin Versfeld.
Subject to the Copyright and Licensing agreements the creators signed with the Computer Science Department at the Univesity of Cape Town.