GazeMatch: Drawing on the Benefits of Motion Matching Gaze Interfaces without Extensive Refactoring of Existing Public Display UIs

Touch controls on public displays, while ubiquitous, introduce public health and accessibility concerns. Alternative controls based on gaze or mid-air interaction techniques usually require a calibration process that hinders spontaneous interaction. Calibration-free gaze alternatives tend to rely on user interfaces (UIs) with moving elements, which require substantial changes to existing UIs, can hinder touch input, and contribute to an overwhelming and confusing user experience -- ultimately limiting adoption. This paper introduces GazeMatch, a novel interaction technique that combines techniques such as Pursuits and Actigaze to address these constraints. We present three user studies that help us define the technique's parameters (N=8), and compare it to state-of-the-art gaze techniques in an abstract selection task (N=12) and in a UI representative of the metro kiosks in New York City (N=5). With little practice, participants committed more errors with GazeMatch than with Google's Look to Speak in the abstract task (19.25% and 4.33%, respectively), but were considerably faster (3.66s and 7.58s, respectively) and ultimately reported preferring GazeMatch. In the representative task, and with no onboarding, participants reported higher discoverability for Pursuits but higher performance for GazeMatch.