logo

Fast Hands-free Writing by Gaze Direction

by David J. Ward and David J C MacKay

In Nature (2002).

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.

Bibtex entry
@article{wardmackay2002,
author="D. J. Ward  and D. J. C. MacKay",
title="Fast Hands-free writing by Gaze Direction",
year="2002",
journal={Nature},
volume={418},
number={6900},
pages={838},
url={http://www.inference.org.uk/mackay/abstracts/eyeshortpaper.html}
}

postscript (Cambridge UK). | postscript (Canada mirror).

pdf (Cambridge UK) | pdf (Canada mirror) | pdf exactly as published in Nature (we prefer our final draft :-)

All postscript files are compressed with gzip - see this page for advice about gzip, if needed.

logo Dasher home page

bibtex entry

@article{wardmackay2002,
author="D. J. Ward  and D. J. C. MacKay",
title="Fast Hands-free writing by Gaze Direction",
year="2002",
journal={Nature},
vol={418},
number={6900},
pages={838},
url={http://www.inference.org.uk/mackay/abstracts/eyeshortpaper.html}
}


related publications.
David MacKay's: home page, publications. bibtex file.
Canadian mirrors: home page, publications. bibtex file.
(c) David J. Ward and David J C MacKay
An Old abstract, longer
We have turned on its head an efficient method for text compression, arithmetic coding, to create an efficient method for text entry. Our text-entry interface, Dasher, is driven by continuous two-dimensional gestures, using a mouse, touch screen, or eyetracker, for example. The user writes by steering through a continuously expanding two-dimensional world containing alternative continuations of the text, arranged alphabetically. A language model predicts which letters might come next and makes the more probable letters easier to write by giving them more space on the screen. The language model can be trained on example documents, and adapts to the user's language as he writes.