Principal Investigator Randall Davis
Project Website http://groups.csail.mit.edu.ezproxy.canberra.edu.au/mug/projects/gesture_kinect/
We developed a unified real-time gesture recognition framework that handles path and pose gestures seamlessly, and responds to discrete and continuous flow gestures promptly and appropriately. The framework is based on hidden Markov models (HMMs), and can take input from both a Kinect sensor and an Inertia Measurement Unit (IMU). Using this framework, we created a browser-based gesture-controlled presentation interface.