A Proposal of Eye Glance Input Interface using Smartphone Built-in Camera

Abstract

The purpose of this study is to develop a line-of-sight input interface using the eye-glance input method, which uses a glance eye movement and method of estimating the eye-control movement by analyzing the image of the area near the eye-ball. The motion was discriminated by taking an image using the built-in camera of a smartphone to rapidly look at the four corners of the smartphone and calculate the moving distance of the eyeball, which was divided into horizontal and vertical components using the optical flow function in Open CV library. In the experiment, we obtained data by asking the subjects to look at the reciprocate of quick gaze movements at each of the four corners and the center of the screen. As a result of data analysis, characteristics of waveforms were obtained for the movements looking at the four corners, and we could determine which of the four corners the users were looking at. Furthermore, by distinguishing between camera shake and line-of-sight movement using the variance value, the discrimination rate was 90% or more even when held in the hand. It was shown that the Eye Glance movement can be measured with a smartphone as well as a PC.