Smile - or the camera will refuse to take your picture

MURDO MACLEOD

CAMERAS never lie, but now they can catch liars. Japanese scientists have developed a camera that is capable of calculating a person's real age and whether their smile is genuine.

The remarkable technology works by comparing the relative size and position of a subject's facial features with a massive database of human faces.

The age calculation can be used in areas such as vending machines and ID theft, and the identification of genuine smiles is potentially useful in mental healthcare, service industry training and even cameras that refuse to take a picture when confronted with a grimace.

The system has been developed by Omron Corporation, based in Kyoto, whose staff analysed about 1.5 million pictures of human faces. They say they started with smiles because they were easier to analyse and are now working on making a computer scan a face for signs of other moods.

First the software confirms it is 'seeing' a face by checking for light and dark shadows cast by different human features. The next step is to scan for clues to the subject's age and mood.

Telltale signs of ageing - at least to a computer - are in the length of the face and the relative distance between the eyes. Younger people have much rounder faces than the elderly, whose faces tend to droop over the years. Age also brings the eyes and eyebrows relatively closer and the nose becomes relatively larger. The number of wrinkles can also be calculated.

The software is being incorporated in cigarette vending machines and as part of extra security in systems to protect against ID theft. Cigarette vending machines are ubiquitous in Japan, and although buyers must use an electronic ID to buy a packet, the age scanner should prevent a young person stealing another person's card and using it to buy tobacco.

The perfect-smile software works by checking key features. In a 'perfect' smile, not only does the mouth change shape but smile wrinkles appear and the eyes appear smaller. In addition, people open their mouths slightly and show their teeth when their smile is genuine.

James Seddon, a spokesman for the firm, said: "It has applications in market testing, for example. If you are testing a new low-fat food and you want to know how much people are enjoying it, then this could help.

"It could also be used in training in the service industry, for example airline staff, so that they can smile at their passengers. Or even robot pets which can react to your mood."

He added: "The [age] technology has a lot of uses, especially in checking a person's identity. In Japan, mobile phones are used to pay for many things and so it is especially important that you can ensure it is not being used by the wrong person."

He admitted that the software would have its limitations. For example, being able to tell the difference between a person who is 19 years and 363 days old - too young to buy cigarettes in Japan - and one who is 20 years and two days old, which is old enough.

This article: http://news.scotsman.com/scitech.cfm?id=1481292007