GazeWheels: Comparing Dwell-time Feedback and Methods for Gaze Input
Misahael Fernandez, Florian Mathis, Mohamed A. Khamis · 2020
We present an evaluation and comparison of GazeWheels: techniques for dwell time gaze input and feedback. In GazeWheel, visual feedback is shown to the user in the form of a wheel that is filled. When completely filled, a selection is made where the user is gazing. We compare three methods for responding to the user when gazing away from the target: Resetting GazeWheel, Pause-and-Resume GazeWheel, and Infinite GazeWheel. We also compare the position of the GazeWheel; Co-located Feedback: shown on the target being gazed at, and Remote Feedback: shown at the top of the interface. To this end, we report on results of a user study (N=19) that investigates the benefits and drawbacks of each method at different dwell times: 500ms, 800ms, and 1000ms. Results show that Infinite GazeWheel and Pause-and-Resume GazeWheel are more error prone but significantly faster than Resetting GazeWheel when using 800-1000 ms dwell time, even when including the time for correcting errors.