People with severe disabilities may soon be able to "train" computers to interpret their individual gestures.
Stephen McKenna and Ian Ricketts of Dundee University's department of applied computing have won Pounds 50,000 from the Engineering and Physical Sciences Research Council for a three-year project to develop "hands-off" computers for people with disabilities.
The award will enable the researchers to work with volunteers from the local McKinnon Centre for the Disabled to develop software that bypasses the traditional keyboard and mouse. Instead, the computer will "see" and respond to the individual's gestures.
Dr McKenna said: "One of the exciting and novel things about this project is that instead of arbitrarily defining a set of gestures that the user has to follow, we are developing a system that can learn, so users can train it to recognise their own individual gestures over a very short period of time."
The project could mean the end of devices that people with disabilities use to operate computers, such as blow tubes and sticks attached to the head.
Dr McKenna has already produced programs that allow computers to "see", enabling images from a digital camera linked to a computer to be simultaneously visualised and interpreted. A computer can, for example, recognise when someone enters a room and identify a face from the crowd.
The new research will expand on this to enable computers to watch, learn, recognise and continue to interpret personal gestures from the hands, face or upper body.
"Furthermore, the software will be able to adapt to changing gestures in order to accommodate users who may have an underlying degenerative disease," said Dr McKenna.
Initially, the project aims to produce a gesture-driven web browser, allowing people to use the web by the mere twitch of a shoulder or any part of the upper body. The Dundee researchers are also working on the related Active project to allow people with disabilities to drive by interacting with a virtual dashboard.
Dr McKenna said the technology could eventually help older people to live independently for as long as possible. Any problems could be monitored through "smart rooms", using a computer to constantly track and interpret activity. "If a person fell and was unable to get up, for example, an alarm would be triggered," he said.