Software that artificially "ages" a person's face using a range of parameters could help find many missing persons, hope the researchers behind it.

The software tries to work out how a face may change by factoring in their personal history, family traits and population trends. Artificial aging software exists already, but this can provide only apply rough physical changes.

Agencies involved in tracing missing persons routinely try to simulate ageing to predict the way a person will look many years after they have disappeared. But they must use artists who are given the most recent images of a person as well as pictures of that person's family.

This is time consuming and expensive, says forensic imaging specialist Chris Solomon at Kent University in the UK: "It can take around 20 to 40 hours per face, which means that generally it is not possible to do as many as would be ideal."

So Solomon and colleagues Catherine Scandrett and Stuart Gibson have developed software that simulates the aging process automatically. It takes into account the way a person has changed in the past, where known, and examines the ageing of other family members as well as the wider population.

Face database

The system first converts a face into a set of numbers based on the location and size of each feature. It then uses a database of previously entered faces to calculate the transformations that need to be made. This database includes previous images of the person in question as well as photographs of their family members and other individuals.

"Most changes in people's faces are shared by the population as a whole," explains Solomon. "We've taken a large sample of faces and extracted the way they change over time, on average."

The ageing software tries to ensure any projected changes bear the right resemblance to images of siblings or parents stored in the database. In an effort to improve the accuracy of the system, the researchers have been using it to artificially age known older people, but using photos of their younger selves. In this way, the accuracy of the technique can be easily gauged.

"The results are generally promising," says Solomon. "Although sometimes it doesn't age a face as much as you would expect." He says these issues are being ironed out.

Numbers game

"It may be useful," says Teri Blythe, who is head of identification and reconstruction at the UK's National Missing Person's Helpline. "There may be trends in how faces change that the software can pick up but a person can't." But Blythe believes the main benefit could be the ability to generate more artificially-aged faces from missing person files.

The National Centre for Missing and Exploited Children, an organisation based in the US, reports that seven out of 10 age-progression images help resolve a case, Blythe notes.

Solomon says the system could also find a range of less serious uses: "More frivolous applications would include sending your friend a birthday card showing what they might look like in another 10 years."