Research

Thesis Title: A real time face recognition system based on a robust and fast face segmentation algorithm using a dynamic template matching approach and a neural classifier

Research Supervisor: Dr. R. D. Sudhaker Samuel

People easily recognize one another by looking at each other's faces. Recognizing human's face is such a fundamental task that even a toddler does it. However, such a trivial task is not simple for computer to perform. A Biometric system is essentially a pattern-recognition system that recognizes a person based on a feature vector derived from a specific physiological or behavioral characteristic of the person. Biometrics represents the most secure way of identifying individuals because verification of identity is established using a physical and behavioural characteristics. To make a personal recognition, biometrics relies on who you are and what you do—as opposed to what you know (such as password) or what you have (such as an ID card). Biometrics has several advantages compared with traditional recognition. In some application, it can either replace or supplement existing technologies, in others it is only viable approach to personal recognition. Several biometric characteristics such fingerprint, face, hand geometry, iris, and voice are in use for various applications. Each biometric has its strengths and weakness, and the choice typically depends on the application. Among all the biometrics, personal recognition based on face is most acceptable since it is nonintrusive. In the recent years considerable progress has been made in the area of face recognition. Computers can now outperform humans in many face recognition tasks, particularly in situation where a face must me searched from a large database.

Face segmentation under Scale variations