Vision & Cognition Laboratory

Department of Computer Science, Drexel University

 
home  |  projects  |  people  |  publications  |  facilities  |  sponsors

Gaze-Based Interfaces

We have also worked toward designing "gaze-driven" systems in which users control a computer using their eyes (possibly with other input). Because of the typical noise and variability in eye-movement data, most existing gaze-driven interfaces incorporate constraints that limit performance in the system (e.g., a minimum "dwell time" that users must spend looking at an item before it activates). In contrast, we have developed two novel gaze-driven interfaces that reduce or eliminate these constraints by using intelligent probabilistic algorithms to interpret eye movements, thus alleviating data noise and increasing usability. One interface, the "eye-typing interface," allows users to type words by looking at letters on an on-screen keyboard [1]. Another interface, IGO, provides a Mac/Windows-like operating-system interface in which users can open and close folders, move icons, and select menu items with gaze input as well as standard mouse input [2]. In both interfaces, users quickly adapt to the new input method and can perform tasks efficiently after very little practice. These interfaces show enormous potential for efficient, user-friendly gaze-driven interaction for both physically disabled and able-bodied users. We are looking into how these algorithms could be integrated into other existing and novel gaze-driven systems.

Primary References

[1] Salvucci, D. D. (1999). Inferring intent in eye-movement interfaces: Tracing user actions with process models. In Human Factors in Computing Systems: CHI 99 Conference Proceedings(pp. 254-261). New York: ACM Press.

[2] Salvucci, D. D., & Anderson, J. R. (2000). Intelligent gaze-added interfaces. In Human Factors in Computing Systems: CHI 2000 Conference Proceedings(pp. 273-280). New York: ACM Press.