Date of Award

Spring 2023

Document Type

Thesis

Publisher

Santa Clara : Santa Clara University, 2023.

Department

Electrical and Computer Engineering

First Advisor

Andy Wolfe

Abstract

Eye tracking is a burgeoning technology that has found applications in market research, accessibility, and as an add-on to games. However, this technology still has room to grow in the consumer space, where touchscreens, keyboards, and mice dominate our interactions with technology. Interfaces that are primarily controlled using eye tracking are less common, and are not found in a handheld form factor. This paper explores the possibility of creating a handheld device that is primarily controlled using eye tracking. Our proposed system performs real-time filtering on eye tracking inputs in order to improve the accuracy and usability of the device. This filtering leverages content-awareness by identifying the interactive elements on the GUI and gravitating the gaze input to those elements. In our single-element speed test, we saw an +11.77% and +11.11% change in speed and accuracy over unfiltered eye tracking, respectively. In our many-element matrix test, we saw a +27.85% and -4.48% change in speed and accuracy over unfiltered eye tracking, respectively. For both tests, eye tracking performed worse in both speed and accuracy than touchscreen or mouse input, both before and after filters were applied. We expect that this disparity could be reduced in the future with better optimization of the system parameters and with more experience using eye tracking as an input.

Download

Share

COinS