Full Professor, University of Calgary, Canada
Title: Machine Learning in KINECT Image Processing for Biometric and Medical Applications
Biography: Marina L Gavrilova
Rapid development of machine learning methods opened the door to a new class of fast and reliable identity management solutions, and forever changed the research landscape. One of the applications where benefits of using machine learning methods are on the full display is in image and video processing. This invited lecture will discuss recent trends, state-of-the-art methods and applications of machine learning in security and medicine. Traditional definition of biometric security research is recognizing someone’s identity from collected biometric data, which includes physiological, behavioural, soft, or social traits. Physiological features can be often collected visually (facial image, ear, iris etc) or through some specialized devices, such as infrared sensors, remote temperature measuring devices, and so on. Behavioural characteristic include the way a person walks (gait), the way person talks (voice), the way person writes (typing patters, keystroke pressure) etc. Soft biometrics include easily collected but not so unique data, i.e. age, gender, height, weight etc. Area where the development of new technologies can have a very tangible effect on society is security and medicine. It is generating a lot of interest and getting traction in biometric research, as well as in related fields looking into human interaction, physiological studies, user profiling, pattern recognition, authorship identification and collective intelligence. This lecture outlines one of the first studies that looks at integrating KINECT sensor image and video processing based on a human gait with activity and emotion recognition. Applications and impact on patient physiohterapy rehabilitation through continuous progress tracking and visual data analytics are also discussed.