Computers recognize the appeal of women

Old grandparents had the saying ' Beauty is in the eyes of the viewers '. But is the person watching it necessarily human?

Tel Aviv University scientists (TAU) think it is not necessarily human. Amit Kagian - master of computer science from TAU school - has succeeded in teaching computer how to interpret women's attraction. But there is a more serious dimension to this problem that is only a fantasy. Discovery is a step towards developing artificial intelligence for computers. Other applications of this software may be in cosmetic and reconstructive surgeries and visual recognition programs on computers such as facial recognition technologies.

From mathematics to aesthetics

'So far, computers have been set up to recognize facial features, such as differences between men and women, and even discover facial expressions,' Kagian said. . 'However, our software allows computers to make esthetic assessments. Associated with abstract thoughts and processes, people can appreciate beauty, but they often don't understand how they came to their conclusions. "

Old grandparents had the saying 'Beauty is in the eyes of the viewers'. But is the person watching it necessarily human? (Photo: iStockphoto)

In the first step of the study, 30 people, both men and women, were shown 100 different faces of Caucasian women (mostly race of white people), most of them were the same age, and these was asked to evaluate the beauty of each face. Subjects evaluated images on a range from 1 to 7, and couldn't explain why they chose certain bases. Kagian and his colleagues then introduced the computer and explained it as well as mapping the geometry of facial features by mathematical method.

Additional characteristics such as the balance of the face, the smoothness of the skin and the smoothness of the hair color are also included in the analysis system. Based on human priorities, the machine 'learned' the relationship between facial features and the bases of attraction, and then tested it on a range of new faces.

Kagian said 'Computers give very impressive results - its classification is similar to the classification that people make . ' Kagian believed that this was seen as an outstanding achievement because the computer "learned" completely how to interpret beauty through the process of processing the data it received earlier.

Beauty is gold

The concept of beauty can be recomposed into binary data and interpreted by a mathematical model is nothing new. More than 2,000 years ago, Greek theologian, philosopher and mathematician Pythagoras observed the connection between mathematics, geometry and beauty. He explained that the characteristics of physical objects correspond to the 'golden ratio' (ideal ratio) is considered the most attractive.

Kagian said 'I know the philosopher Plato binds good to beauty. I personally believe that some of the world's authenticity about beauty exists in nature is the aesthetic interpretation of global obviousness. However, because each of us is stuck with personal prejudices and personalized views, this can distract us in finding the final formula to understand. absolutely about beauty. '

Kagian - who studied under the direction of TAU's multidisciplinary research program for excellent students - said that the next step would be to teach computers how to identify men's 'beauty' . . This may be more difficult. Psychological research shows that there is very little general agreement about the definition of "male beauty" among human subjects. And Kagian joked that his own portrait would not be part of the experiment.

He said 'Maybe I will blow up that machine.'

Kagian published these findings in the scientific journal Vision Research. The co-authors studying this work are two supervisors of Kagian: Professor Eytan Ruppin and Professor Gideon Dror.