A longitudinal study of text entry by gazing and smiling

Abstract

Face Interface is a wearable device that combines the use of voluntary gaze direction and facial activations, for pointing and selecting objects on a computer screen, respectively. In this thesis a longitudinal study for entering text using Face Interface is presented. The aim of the study was to investigate entering text with Face Interface within a longer period of time. Twelve voluntary participants took part in an experiment that consisted of ten 15-minutes long sessions. The task of the participant in each session was to write text in fifteen minutes with Face Interface and an onscreen keyboard. Writing was done by pointing at the characters by gaze and selected by smiling. The results showed that the overall mean text entry rate for all sessions was 5.39 words per minute (wpm). In the first session the overall mean text entry rate was 3.88 wpm, and in the tenth session 6.59 wpm. The overall mean minimum string distance (MSD) error rate for all sessions was 0.25. In the first session the overall mean MSD error rate was 0.50 and in the tenth session 0.05. The overall mean keystrokes per character (KSPC) value for all sessions was 1.18. In the first session the overall mean KSPC value was 1.26 and in the tenth session 1.2. Subjective ratings showed that Face Interface was easy to use. The rating of the overall operation of Face Interface was 5.9/7.0 in the tenth session. Subjective ratings were positive in all categories in the tenth session.