Real-time stereo face tracking system for visual human interfaces

When a person instructs operations to a robot, or performs a cooperative task with a robot, it is necessary to inform the robot of the person´s intention and attention. Since the motion of a person´s face and the direction of the gaze is deeply related with person´s intention and attention, detection of such motions can be utilized as a natural way of communication in human-robot interaction. In this paper, we describe our real-time stereo face tracking system. The key of our system is the use of a stereo vision. Since the 3D coordinates of the features on the face can be directly measured in our system, we can drastically simplify the algorithm of the 3D model fitting to obtain the full 3D pose of the head compared with conventional system with monocular camera. Consequently we achieved a non-contact, passive, real-time, robust, accurate and compact measurement system of human´s head pose and gaze direction