Security Based on Real Time Tracking of Multiple Human Faces Identification

Robust tracking of persons in real-world environments and in real-time is a common goal in many video applications. In this paper a computational system for the real-time tracking of multiple persons in natural environments is presented. Face detection has diverse applications especially as an identification solution which can meet the crying needs in security areas. The region extractor is based on the integration of skin-color, motion and silhouette features, while the face detector uses a simple, rule-based face detection algorithm and SVM. Exemplary results of the integrated system working in real-world video sequences. New intelligent processing methods, as well as security requirements make multiple-person tracking a hot area. This application is robust tracking in real-world environments and in real-time.