Implementation of a Mobile Device Eye-tracking System Using Augmented Reality

Abstract: Eye-tracking technology is used within various fields of research such as medicine, computer vision and psychology among others. Despite the many application areas and a growing interest, most eye-trackers rely on dedicated and expensive hardware to track eye movements on large monitors such as desktop displays. Eye-tracking on mobile devices using their built-in camera would be beneficial on a large scale since additional, expensive, equipment would not be needed. Because of the many smartphone users, the technology would be more available. This study aimed to explore the possibilities of using a smartphone camera and augmented reality (AR) for eye-tracking, as well as evaluate the accuracy and precision such a system could obtain. The methods used for gaze estimation were a ray-plane intersection and a system-controlled calibration using linear scaling and translation to adjust the future gaze positions. The mobile device eye-tracker was evaluated in terms of accuracy and precision by two validations using a point pattern to collect gaze data on fixations detected by the system. The result showed an accuracy and precision below optimal compared to previous studies on mobile eye- trackers and eye-trackers that use infra-red light to track the eye-movements. Several factors could have affected the result such as the environments in which the evaluations were held, eye-blinks during data collection or the implementation of the eye-tracking system including the method for calibration. Future work is recommended to evaluate the system in a controlled lab environment, including a larger number of participants and examine if the accuracy is mainly affected by external factors or if it is affected by the implementation of the system. Because of the many benefits of mobile device eye- tracking, future work within the area is encouraged. It would be interesting to investigate the ability to use a mobile device eye-tracker with unrestricted head- and hands movement for further use-cases and a more natural experience.