Towards gaze analysis in mobile applications

Gaze analysis gives us detailed information on the visual attention of a person. Further insights into the world of thought can be revealed and the information can contribute to interest or intention deduction. That is why eye tracking is extremely interesting for a variety of applications and different research fields. Especially due to the usage of mobile unobtrusive eye trackers, experimental setups can stay closer to reality and eye tracking becomes possible in settings where one didn't imagine it to be realisable years ago. But still, commercial off-the-shelf eye trackers do neither enable for 3D gaze point computation nor have solutions for fully automated gaze analysis in environments with real 3D objects. In this article we show which information mobile eye trackers need to deliver to make 3D gaze point computation on real 3D objects in known environments possible. We also demonstrate how this has been achieved for the Dikablis Wireless eye tracking system. We further show real time processing of 3D gaze points for fixation determination and an accuracy test to evaluate the developed algorithms.