Facial component detection using a Kinect for an Interactive Media Art

The 3D face extration and facial component detection are important preprocessing step of automatic human recongition for an interactive system. Using proint cloud from a depth sensor, the human and backgrouned are seperated using depth information. The head and body are seperated using the geometric characteristics of human shape. After finding nose tip using depth information of head, nose, forhead, eyes and lips are classified as facial components using depth curves in xand y-coordination based on the position of nose tip.