Triangle similarity approach for detecting eyeball movement

Abstract

Eye movement detection is one of the most developed technologies in the field of human and computer interaction especially as a control instrument of an automatic device. Eye movement of user can provide an easy input source, natural and high-bandwidth. By tracking the user's point of view, the ease of communication from the user to the automatic device can be improved. In general, the process of tracking the eye movement involves the process of detecting the eye area first. However, unlike the methods development in the eye detection process, methods development to detect eye movements still need to be improved. In this paper, we used a triangle similarity formulation to derive angular values by calculating the angle of a line between two tracked points obtained from face and eyeball midpoint, relative to the horizontal. The angle value and the length of the slash line are used as an indicator of the eyeball movement. This approach can accommodate various directions of movement of the eye even produce accuracy and precision well enough evidenced by the percentage of detection success reached 79%. This approach can also be used as a possible alternative way to detect the eyeball movements.