EyeDroid
Shahram Jalaliniya, Diako Mardanbegi, Ioannis Sintos, Daniel Garcia Garcia · 2015
In this paper we report on development and evaluation of a video-based mobile gaze tracker for eyewear computers. Unlike most of the previous work, our system performs all its processing workload on an Android device and sends the coordinates of the gaze point to an eyewear device through wireless connection. We propose a lightweight software architecture for Android to increase the efficiency of image processing needed for eye tracking. The evaluation of the system indicated an accuracy of 1.06 degrees and a battery lifetime of approximate 4.5 hours.