Improved multi-camera 3D Eye Tracking for human-computer interface

A human being accumulates the majority of the surrounding impressions through vision. With the ability to track and observe a user´s gaze, it is possible to obtain deep insights into the person´s attention and, based on the gathered information, understand his or hers behavior and actions. In this paper we present an improved Multi-Camera 3D Eye Tracking framework for human-computer interface based on an algorithm that provides automated control of a Pan-Tilt camera (PTC) to follow a person´s eye. We employ a fixed 3D wide-angle depth sensor (Kinect) in order to determine the position of the human face and its features, most importantly the eyes. By means of the supervised descent method (SDM), we compute correctly the position of the two eyes using 6 landmarks for each eye and the position of the head. Then, an active pan-tilt camera is oriented to one of the users´ eyes, and in this way, a high-precision eye tracking is accomplished.