Fast Hands-free Writing by Gaze Direction

We describe a method for text entry based on inverse arithmetic coding that relies on gaze direction and which is faster and more accurate than using an on-screen keyboard. These benefits are derived from two innovations: the writing task is matched to the capabilities of the eye, and a language model is used to make predictable words and phrases easier to write. For people who cannot use a standard keyboard or mouse, the direction of gaze is one of the few ways of conveying information to a computer. Many systems for gaze-controlled text entry provide an onscreen keyboard whose buttons are ‘pressed’ by staring at them. But eyes did not evolve to push buttons, and this method of writing is exhausting. Moreover, on-screen keyboards are inefficient because typical text has considerable redundancy.1 Although a partial solution to this defect is to include word-completion buttons as alternative buttons alongside the keyboard, a language model’s predictions can be better integrated into the writing process. By inverting an efficient method for text compression – arithmetic coding2 – we have created an efficient method for text entry, one that is also well matched to the eye’s natural talent for search and navigation. One way to write a piece of text is to go into the library that contains all possible books,3 and find the book that contains exactly that text. Writing thus becomes a navigational task. In our idealized library, the ‘books’ are arranged alphabetically on one enormous shelf. As soon as the user looks at a part of the shelf, the view zooms in continuously on the point of gaze. To write a message that begins ‘hello’, one first steers towards the section of the shelf marked h, where all the books beginning with h are found. Within this section are sections for books beginning ha, hb, hc, etc.; one enters the he section, then the hel section within it, and so forth. To make the writing process efficient we use a language model, which predicts the probability of each letter’s occurring in a given context, to allocate the shelf-space for each letter of the alphabet (figure 1a). When the language model’s predictions are accurate, many successive characters can be selected by a single gesture. We previously evaluated this system, which we call Dasher, with a mouse as the steering device.4 Novices rapidly learned to write and an expert could write at 34 words per minute. All users made fewer errors than when using a standard qwerty keyboard. Figure 1b shows an evaluation of Dasher driven by an eyetracker, compared with an on-screen keyboard. After an hour of practice, Dasher users could write at up to 25 words per minute, whereas on-