Emotionally Aware Computers to Help Fight Autism

The visitors to Royal Society summer science exhibition may help teach computers how to read facial expressions by taking part in research with computers designed to mind-read by analyzing facial expressions. Hopefully, the research will lead to the development of ways to assist people with autism to recognize emotions.

Peter Robinson, professor of computer technology at the University of Cambridge, said, “Imagine a computer that could pick the right emotional moment to try to sell you something, a future in which mobile telephones, cars and websites could read our minds and react to our moods.”

The developers of this technology are working with colleagues in America to develop a headset version of the system that would interpret others’ moods and communicate them to the wearer. The headset would be a great help for people who find it difficult to read others’ facial expressions and emotions, similar to those with autism and Asperger’s syndrome.

The prototype, developed at the university, will be unveiled at the summer science exhibition in London.

In this computer program, data from a camera locates and tracks 24 facial “feature points”, such as the edge of the nose, the eyebrows and the corners of the mouth. So far, 20 key facial movements have been connected to respective emotions.

There are small variations in the way people express the same emotion. The system is being refined by testing it with “real” people’s expressions to everyday life using cameras attached to neck-braces. The exhibition is expected to generate valuable new data to improve the program’s ability to read faces. Professor Robinson said, “The system can cope with the variation in people’s facial composition.”

Researches are also being conducted for using this technology in cars, as it can improve driver safety. “We are working with a major car company and it is possible that this technology could feature in cars within five years,” said Professor Robinson. The research team is recording the facial expression of volunteers in various driving conditions and monitoring facial movements to identify more complex expressions linked to confusion, boredom or tiredness.