Gaze estimation using Kinect/PTZ camera

Abstract

This paper describes a novel method for eye-gaze estimation under normal head movement. In this method, head position and orientation are acquired by Kinect while eye direction is obtained by PTZ camera. We propose the Bayesian multinomial logistic regression based on a variational approximation to construct a gaze mapping function from head and eyes features. Our proposed method eliminates stationary head position, awkward personal calibration procedure and active light source as three common drawbacks in most conventional techniques. The efficiency of the proposed method is validated by performance evaluation for different users under varying head position and orientation.