Eyes Tracking in a Video Sequence Based-On Haar-like Features and Skin Color

Abstract

This paper presents a method of face and eyes locating and tracking in a video sequence based on haar-like feature and face skin color. This method first extracts the haar-like feature of face and utilize adaboost algorithm to detect face in image. Then a appropriate skin-model is employed to dispose and locate the eyes based on the geometrical model. Finally the face and eyes tracking in the video sequence is achieved. Testing with a general PC camera in the laboratory environment shows that the method provides high precision and speed.