Eye-gaze estimation under various head positions and iris states

作者：

Highlights：

• A new method for eye-gaze estimation under normal head movement is investigated.

• Head position and orientations are acquired by Kinect depth data and eye direction is obtained from high resolution images.

• Bayesian multinomial logistic regression is used to construct a gaze mapping function and to verify iris state.

• The features used for learning, vary with different head positions and orientations and eye directions.

• Whenever the user moves away from the camera, the gaze score will decrease.

摘要

•A new method for eye-gaze estimation under normal head movement is investigated.•Head position and orientations are acquired by Kinect depth data and eye direction is obtained from high resolution images.•Bayesian multinomial logistic regression is used to construct a gaze mapping function and to verify iris state.•The features used for learning, vary with different head positions and orientations and eye directions.•Whenever the user moves away from the camera, the gaze score will decrease.

论文关键词：Gaze estimation,Bayesian logistic regression,Kinect

论文评审过程：Available online 22 August 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.08.003