Facial Orientation and Eye-Gaze Detection Based on Real-Time 3D Facial Imaging By Using Correlation Image Sensor

3D facial orientation and eye-gaze detection is one hot area in human-machine interaction and application, it is also a challenging field in computer vision, pattern recognition and image sensing. We designed a real-time 3D facial orientation and eye-gaze detection system based on 3D real-time facial imaging system by using three-phase correlation image sensor. Based on the normal vector map and intensity image obtained by our imaging system simultaneously, the differential geometry theory is employed to find the key points of face image, according to these key points and curvature feature of face, the important eye area could be segmented from the face image. The optical flow algorithm originally presented by Bruce D. Lucas and Takeo Kanade is implemented to get the detection result. The experiments show that the performance for our test is satisfying