The Measure of False Beauty

Last summer, Julia Roberts made headlines with “Too pretty to be true” and it was more than tabloid headlines. The British Advertising Standards Authority had banned a campaign by the L’Oréal cosmetics company. The reason: Julia Roberts, a forty-three year-old mother of three, just looked too perfect on the photos. The advertisers had done a lot of digital manipulation and banished blemishes and wrinkles from the Hollywood star’s face.

Ms.Roberts is only one of many in a long list of stars and models who have been embellished beyond recognition by digital image retouching. In the advertising industry excessive retouching is taken for granted. The problem is: magazine readers cannot distinguish to what extent photos have been manipulated. They compare themselves to the unnatural beauties and feel pressured. But this is about to change.

Computer scientists Eric Kee and Hany Farid from Dartmouth College in New Hampshire want to help frustrated consumers. They have developed a computer program that automatically recognizes and evaluates digital alterations in photos. Their goal is to ease the way for the labeling of retouched photos in magazines, a requirement that the advertisement industry has successfully refused to agree to. “Thanks to the magic of digital retouching, impossibly thin, tall, and wrinkle-free models routinely grace advertisements and magazine covers with the legitimate goal of selling a product to consumers. On the other hand, an overwhelming body of literature has established a link between idealized and unattainable images of physical beauty and serious health and body image issues for men, women, and children. It is our hope that a perceptually relevant metric of photo retouching can help find a balance between these competing interests,” said the scientists in their article published in the Proceedings of the National Academy of Sciences

The crowd also played a significant part in the first steps towards the creation of the program. In order to give the program an initial benchmark the Internet users evaluated a series of photos by hand. Their task was to evaluate 20 pairs of original photos and retouched images according to the extent of alterations and classify them on a scale of one to five. This gave the scientists 50 evaluations for each of the 468 pairs of photos; data they were able to use for the further development of the program. After a series of calculations and a lot of programming, they created a program that automatically recognizes the advertisers tricks and might possibly contribute to the worldwide debate.

To find out more, read the full article published in the Proceedings of the National Academy of Sciences