부산대학교 도서관

This paper presents an integrated security management for authentication of users based on dynamic biometric features. For the vary applications of mobile-commerce, a mobile device cooperating with a user calibration interface is used to capture the dynamic face images of the sliding view of the human face to improve the security level of the whole system. The dynamic face recognition automatically chooses proper face image from the sliding view according to the normal vector of the human face. It can provide high face recognition rate from different viewing angle and reduce computation time to fit the requirement of the platform on smart phones. The adaptive model for detection deals with the variant lighting conditions. Then, the weighted eigenface integrates features from different viewing angles between the viewing direction and the normal vector. The experimental results showed that the proposed method can be successfully applied to the smart phone for authentication of users under variant environments.