A low complexity face recognition for consumer devices

Due to the rapid growth of consumer grade devices and corresponding application market, the incorporation of vision capabilities into embedded systems has gained significant attention from researchers lately. Similarly to the human visual system, embedded computer vision systems analyze and extract information from visual content in a wide variety of products. Face recognition has been one of the most successful applications in this field. A cost effective implementation of reliable face recognition (FR) solutions can be useful for a wide range of applications, such as identity authentication, entertainment, and content based retrieval system. However, embedded system based face recognition solutions often suffer from not only common problems such as variation in illumination (pose variation does not appear to be a problem for face recognition at a short distance), but also relatively low quality input face images, and limited computational resources. Although many researchers have attempted to develop robust FR algorithms, relatively few initiatives have been undertaken to adapt FR solutions to mobile/smart-phone market. It is believed that the research results obtained by this work will strengthen our industrial partners’ technological leadership and competitiveness.

Faculty Supervisor:

Dr. Konstantinos Plataniotis

Student:

Wang Jeaff Zheng

Partner:

Qualcomm Canada Inc.

Discipline:

Engineering - computer / electrical

Sector:

Information and communications technologies

University:

University of Toronto