Real Time Face Selection and Tracking Based Camera Motion Control System

This paper describes the design and implementation of a architecture that can continously monitor and track selected face in real time by controlling camera motion around. Dection and tracking of face has been an significant and dynamic investigation domain because it provides many application, primarily in video coding, video surveillance, biometry or Video Identity Resolution. But detection and tracking of face in real time is a composite issue with numerous feasible applications. The objective of this project is to implement an real time camera motion control system to detect and track selected human face. The detection of face includes cascade classifiers and harr features. One of the detected face will be selected using mouse click option and matching is done in each frame by difference algorithm. The system is integrated with a frame-by-frame Kalman tracker in order to locate face region in each frame. Experimental result demostrate that the system able to track selected face in real time with horizontal rotation angle -360 to +360 degree with maximum rotation speed 5 degree per second. Implementation was done in real time with minimum computational endevour, thus suitable for low cost application.