A real-time security surveillance system for personal authentication

Abstract

We introduce a human surveillance system, which integrated technologies of face tracking, face detection, and face recognition to authenticate personal identity in real time. The system located all of the faces in the view quickly and tracked them all the time even though people crossed. Facial features of a selected target were estimated, and the personal identity was justified. In our designation, spatial and temporal information over a number of continuous frames were accumulated and combined to make the system more stable and robust to unexpected human activities. More than 60 persons of an experiment were trained and recognized in the natural environment. Each of them was effectively authenticated by the system in terms of recognition accuracy and performance.