Gaze Supported Gestural Computer Interaction: Performance Implications of Training

Abstract

Gaze supported non-tactile gestural control uses a combination of gestures based body movements with eye gaze positioning to provide an input source for a user’s control with a system. Combining body gestures with eye movements allows for unique computer control methods other than the traditional mouse. However, research is mixed on the effectiveness of emerging control types, such as gestures and eye-tracking, with some showing positive performance outcomes for one or more control aspects but performance detriments in other areas that would prohibit the use of such novel control methods. One important aspect that is often ignored is familiarity with the control method. Unlike the mouse, users are typically unfamiliar with eye and gestures based control methods. In order to truly understand the benefit of new concepts like gaze supported gestural controls, testing experienced users is necessary. In the current experiment, participants were trained on the gaze supported gestures system in order to become “experts” and achieve similar levels of proficiency with the different control methods to be assessed, to include mouse, non-gaze and gaze supported gestural controls. Results showed that after as few as five practice sessions participants were able to perform a simple point and click task as well or even better than mouse control when using gaze supported gestures.