If a bunch of researchers from Tel Aviv University get to have their way, soon, computers (in addition to men) will be chasing skirts. The researchers claim they've managed to successfully 'teach' a computer -- any computer how to spot an attractive woman.

The project's leading light, an Indian Israeli by the name Amit Kagian, along with colleagues Professor Eytan Ruppin and Professor Gideon Dror explained that they'd married the worlds of computer programming and psychology to get a computer to make an aesthetic judgement. Said Kagian, "Until now, computers have been taught how to identify basic facial characteristics. But our software lets a computer make an aesthetic judgment."

The first step of the study had 30 men and women judging the beauty of 100 different faces of Caucasian women, roughly the same age. The men and women rated the faces for beauty on a scale of one to seven, not giving reasons for the scores they meted out. The second step saw Kagian and team processing, mapping geometrical shapes of facial features mathematically into the computer. The researchers also fed features like face symmetry, skin smoothness, and hair color into the analysis. It was found that based on human preferences, the computer 'learnt' the relationship between facial features and attractiveness quotient (scores). This ability of the computer was subsequently put to test on a fresh set of faces. The third and final step of the study had a jubilant Kagian say, "The computer produced impressive results -- its rankings were very similar to the rankings that people gave." Kaigan said this is indeed a remarkable achievement for its as though the computer had 'learnt' implicitly how to interpret beauty through processing previous data it had received.

And while beauty may be skin deep, on a more serious note, the study is significant in the development of artificial intelligence (AI) in computers. It may see application in fields such as plastic and reconstructive surgery, and computer visualization programs like face-recognition technologies.