Development of ET-DR Algorithm to Enhance Resolution for Gaze Correction of Low-Resolution Image Based Webcam Eye Tracking

Abstract

Eye tracking is method of detection of a point on a screen that the user gazes at on CV technology. Eye tracking technology is one of the NUI technologies which enables controlling of GUI or inputting text, but it is also possible to analyze the user’s psychological state or the user’s understanding of the contents being gazed at through the user’s eye gaze analysis. Eye tracking can be used for reading education, internet shopping, commercial advertising, etc. Furthermore, it can be used to detect cheating in online exams. The eye tracking system’s permissible range varies depending not only on the motion of the user, but also on the camera, the quality of the acquired image, and the difficulty of the algorithm. Therefore, the accuracy of eye tracking is greatly influenced by how the eye tracking algorithm is implemented in addition to hardware variables. This paper presents Eye tracking Dead Reckoning (ET-DR) as a method to increase the resolution of the eye-tracking system for reading education. DR is usually an algorithm used in navigation, but it has been modified and optimized to increase the resolution of eye tracking. To verify this algorithm, 100 Korean elementary school kids read Korean texts. At the time, the eye tracking success rate was assessed using a computer with an ET-DR algorithm and a computer without the algorithm, and a highly relevant result was achieved.