Pose determination and viewpoint determination of human head in video conferencing based on head movement

The lack of eye contact in video conference degrades the user experience. This problem has been known and studied for many years. To correct the eye contact, it requires face-reorientation and eye gaze correction. Pose determination and viewpoint determination is a vital step prior re-orientating the face. This paper suggests a novel, simple and effective approach for the determination of pose and viewpoint based on head movement. Our approach is designed for the video-based situations, that it takes the advantage of the availability of a previous image, and initially with a known pose or viewpoint. It needs one feature point to determine the pitch angle and the yaw angle, and two feature points to determine the roll angle. The number of feature points to be tracked is significantly less than many other researches suggested in this area. Our approach has little overhead to the face re-orientation process. It is fast and efficient, and produces reasonable accuracy.