TOAST

Abstract

Touch typing on flat surfaces (e.g. interactive tabletop) is challenging due to lack of tactile feedback and hand drifting. In this paper, we present TOAST, an eyes-free keyboard technique for enabling efficient touch typing on touch-sensitive surfaces. We first formalized the problem of keyboard parameter (e.g. location and size) estimation based on users' typing data. Through a user study, we then examined users' eyes-free touch typing behavior on an interactive tabletop with only asterisk feedback. We fitted the keyboard model to the typing data, results suggested that the model parameters (keyboard location and size) changed not only between different users, but also within the same user along with time. Based on the results, we proposed a Markov-Bayesian algorithm for input prediction, which considers the relative location between successive touch points within each hand respectively. Simulation results showed that based on the pooled data from all users, this model improved the top-1 accuracy of the classical statistical decoding algorithm from 86.2% to 92.1%. In a second user study, we further improved TOAST with dynamical model parameter adaptation, and evaluated users' text entry performance with TOAST using realistic text entry tasks. Participants reached a pick-up speed of 41.4 WPM with a character-level error rate of 0.6%. And with less than 10 minutes of practice, they reached 44.6 WPM without sacrificing accuracy. Participants' subjective feedback also indicated that TOAST offered a natural and efficient typing experience.