Case Study: Improving gaze typing for the disabled

Background image

Researchers from Microsoft and University of Washington Seattle used SMI Eye Tracking to come up with a better method of hands-free, gaze typing.

Challenge

For many disabled people the ability to type with their eyes – gaze typing – transforms their interaction with the world. But how can researchers help them by improving typing speeds and minimizing errors? Researchers from Microsoft and the University of Washington Seattle came together in this study to find out.

Solution

This research focused on dwell time – the time set by the gaze typing software required to select a target key or letter. If the dwell time is set too long, the typing speed becomes too slow but if it is too short, then the number of errors increases.

The solution is cascading dwell time – a system that dynamically adjusts the dwell time of keys in an on-screen keyboard based on the likelihood that a key will be selected next, and the location of the key on the keyboard.

SMI Eye Tracking

Researchers used an SMI REDn Eye Tracker connected to a Microsoft Surface Pro 3 tablet with a 12-inch (30cm) screen. In this case the eye tracking didn't just select the keys, but supplied data essential to the setting the dwell times within the software.

Benefit

This system gave the users a smoother experience and a faster typing speed with fewer errors. With cascading dwell time, users achieved average speeds of 12.39 words per minute (wpm) whereas the average with static dwell time was 10.62.

The corrected error rate was also lower with the cascading system – 9.63 per cent as opposed to 14.88 with static dwell time. The study concluded: "This work takes a significant step toward improving gaze base text entry, a life-changing form of interaction for many people with motor disabilities."

Background

For people with neurodegenerative diseases or severe motor disabilities, interaction with computing devices may be limited to eye gaze, as their mobility restrictions can make a mouse, keyboard, or touch screen difficult or impossible to operate. As a result, text entry is often accomplished through gaze typing achieved through the use of an eye tracking device and an on-screen keyboard.

Benefit

This system gave the users a smoother experience and a faster typing speed with fewer errors. With cascading dwell time, users achieved average speeds of 12.39 words per minute (wpm) whereas the average with static dwell time was 10.62.

The corrected error rate was also lower with the cascading system – 9.63 per cent as opposed to 14.88 with static dwell time. The study concluded: "This work takes a significant step toward improving gaze base text entry, a life-changing form of interaction for many people with motor disabilities."

Customer institution University of Washington, Microsoft Corporation
Customer website https://www.washington.edu/
Related publication