Multi Gaze Interactions application provides multi-user gaze tracking, which can be used with any standard camera in real-time. The application includes a pre-trained model using Columbia Gaze Dataset for gaze prediction and Haar cascade filters for face and facial features detection.
The Multi Gaze Interactions was written solely in Python3. Therefore, if the users have Python installed already on the machine, it is a matter of a couple of command lines the Multi Gaze Interactions application is ready to be used. To run the application, the users need to connect their webcams and enter a command for the Python code to run.
The users can run the Multi Gaze Interactions Python application from any machine. However, there are a couple of requirements that need to be met. The application was implemented using Python3, and it requires the machine to have Python 3.7 or higher installed. It also requires 'pip' version 19.3 or higher.
In the current version we use a heuristic driven approach for face, facial features and gaze detection. The source includes pretrained model using Columbia Gaze Dataset for gaze prediction and Haar cascade filters for face and facial feature detection.