Grade Level at Time of Presentation
Senior
Major
Computer Information Technology
Minor
Music
2nd Grade Level at Time of Presentation
Sophomore
2nd Student Major
Computer Science, Applied software engineering
Institution
Northern Kentucky University
Faculty Advisor/ Mentor
Dr. Nicholas Caporusso
Department
Computer Science
Abstract
Gaze tracking has become an established technology that enables using an individual’s gaze as an input signal to support a variety of applications in the context of Human-Computer Interaction. Gaze tracking primarily relies on sensing devices such as infrared (IR) cameras. Nevertheless, in the recent years, several attempts have been realized at detecting gaze by acquiring and processing images acquired from standard RGB cameras. Nowadays, there are only a few publicly available open-source libraries and they have not been tested extensively. In this paper, we present the result of a comparative analysis that studied a commercial eye-tracking device using IR sensors, that is, Tobii 4C and WebGazer, a software system that uses machine learning and linear regression to estimate gaze from images acquired by a standard webcam. From our findings, we can conclude that, despite the advancements in artificial intelligence and computer vision, gaze tracking using IR sensors is still significantly more accurate than RGB webcams. Specifically, the software library tested in our work is not suitable for gaze tracking tasks that require accuracy and reliability.
Included in
Artificial Intelligence and Robotics Commons, Digital Communications and Networking Commons, Graphics and Human Computer Interfaces Commons, Hardware Systems Commons, Signal Processing Commons, Software Engineering Commons
Comparative Analysis of RGB-based Eye-Tracking for Large-Scale Human-Machine Applications
Gaze tracking has become an established technology that enables using an individual’s gaze as an input signal to support a variety of applications in the context of Human-Computer Interaction. Gaze tracking primarily relies on sensing devices such as infrared (IR) cameras. Nevertheless, in the recent years, several attempts have been realized at detecting gaze by acquiring and processing images acquired from standard RGB cameras. Nowadays, there are only a few publicly available open-source libraries and they have not been tested extensively. In this paper, we present the result of a comparative analysis that studied a commercial eye-tracking device using IR sensors, that is, Tobii 4C and WebGazer, a software system that uses machine learning and linear regression to estimate gaze from images acquired by a standard webcam. From our findings, we can conclude that, despite the advancements in artificial intelligence and computer vision, gaze tracking using IR sensors is still significantly more accurate than RGB webcams. Specifically, the software library tested in our work is not suitable for gaze tracking tasks that require accuracy and reliability.