Abstract
Eye gaze tracking plays an important role in human-computer interaction applications. In recent years, many research have been performed to explore gaze estimation methods to handle free-head movement, most of which focused on gaze direction estimation. Gaze point estimation on the screen is another important application. In this paper, we proposed a two-step training network, called GazeEstimator, to improve the estimation accuracy of gaze location on mobile devices. The first step is to train an eye landmarks localization network on 300W-LP dataset [1], and the second step is to train a gaze estimation network on GazeCapture dataset [2]. Some processing operations are performed between the two networks for data cleaning. The first network is able to localize eye precisely on the image, while the gaze estimation network use only eye images and eye grids as inputs, and it is robust to facial expressions and occlusion.Compared with state-of-the-art gaze estimation method, iTracker, our proposed deep network achieves higher accuracy and is able to estimate gaze location even in the condition that the full face cannot be detected.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
