Israeli photo application promises more beautiful you

Since its unveiling in Boston, the response to the Beauty Function has been overwhelming: Media, including New Scientist and Forbes, have been eager to report on a computer program that can change the landscape of digital photography.

[additional-authors]

Karin Kloosterman

February 15, 2007

If the camera could lie, would you let it?

Three Israeli computer scientists from Tel Aviv University have developed the ultimate enhancement tool for retouching digital images. Called the Beauty Function, their program scans an image of your face, studies it and produces a slightly more beautiful you.

Introduced at a conference in Boston recently after more than three years of work, the Beauty Function is the inspiration of Tel Aviv University’s Daniel Cohen-Or and Tommer Leyvand.

In developing the Beauty Function, they asked 300 men and women to rank pictures of peoples’ faces — with varying degrees of beauty — on an attractiveness scale of 1-7. The scores were correlated to detailed measurements and ratios of facial features, such as nose width, chin length and distance from eyes to ears.

Some 250 measurement points were taken into account and, once formulated, researchers developed an algorithm that let them apply some of the desired elements of attractiveness — as mathematical equations — to a fresh image.

The result is a computer program that within minutes can decide how to make you more beautiful. Larger eyes perhaps? A less-crooked nose? How about lips slightly closer to the chin?

When carried out on a large number of sample images, volunteers agreed that 79 percent of time the effects of the Beauty Function — which can be applied to both men and women — made a face more attractive.

Photo-editing software companies such as Adobe (manufacturer of Photoshop) are potential customers of the new tool, and researchers hope it will also become a must-have add-on for all digital cameras in the future, “just like the red-eye function is today,” Leyvand said.

Like a true scientist, Leyvand has also tried using the Beauty Function on himself and family members. One relative told him that she was pleased with the output.

“She told me, ‘Now I know what I need to do to improve my makeup application,'” Leyvand said.

“If you can understand what the algorithm of the Beauty Function has chosen to do on your face,” he added, “it can help you accentuate parts of yourself deemed more attractive. You might want to use more lipstick to make your lips fuller.”

Plastic surgeons, he adds, may find it helpful to increase business. With a flick of a switch they can show people how minor alterations on the face and neck can enhance attractiveness.

Chances are most people will opt to keep enhancements in the realm of the digital world. And there is a need: It is no big secret that celebrities and models are being digitally enhanced in pictures and magazines. Why shouldn’t all of us enjoy some of that picture-perfect retouching too?

“Beauty is not in the eye of the beholder,” co-researcher Cohen-Or said. “Beauty is merely a function of mathematical distances or ratios. And interestingly, it is usually the average distances to features which appears to most people to be the most beautiful.”

“I don’t know much about beauty and I don’t pretend that I do,” he added, “but the nice thing about this project is that we didn’t intend or aim to define beauty. We don’t care about the reasons that make someone appear to be more beautiful. For us, every picture is just a collection of numbers.”

Leyvand and Cohen-Or envision that such a tool will be used for producing the ultimate dating site picture, and as a one-stop-shop enhancement tool for photo editors at glossy magazines.

Whatever the purpose behind using Beauty Function, the researchers are confident it will make a splash in the photo-editing world. Unlike existing software that relies on human intervention to decide what changes to make, the Beauty Function uses the computer to decide. Also, current touch-up software has magazine editors complaining of doctored images looking “cartoony” and little like the original. By comparison, the output of the Beauty Function looks natural.

The Beauty Function idea started around the time Leyvand had finished his master’s degree in 2003. Lingering around the computer science lab at Tel Aviv University, he continued to ping-pong ideas off his former mentor, Cohen-Or. Together they decided to build on a body of work in the area of computer learning, which was started by Dr. Gideon Dror at the Academic College of Tel-Aviv-Yaffo.

“When I thought about what he did, I thought about using his idea to guide an actual change towards making a picture more beautiful,” Leyvand recalled.
Today, Leyvand is in Redmond, Wash., working for Microsoft as a computer developer, while Cohen-Or has taken on the task of commercializing the beauty software.

As part of his ongoing work as a computer scientist, Cohen-Or also works with the notion of finding a similar beauty function related to color. Color harmonies exist, he said, yet not a lot has been done with aesthetics and color. Finding or matching the right harmonies of color — opposites or colors belonging to the same hue — can have a big impact on advertising and art, he believes.

But with or without color, the Beauty Function is bound to impact the way snapshots of our faces are taken and processed.

“Think about how great this could be for a professional photographer at a photo shoot,” Leyvand said. “Normally they take hundreds of pictures to capture the right expression for the perfect shot. It is a rare combination of light, camera position and angle of the face that makes the perfect picture.

“Getting that moment is a kind of magic. I think with our software we can capture that magic moment every single time.”

Tel Aviv University Homepage:
‘ TARGET=’_blank’>http://www.cs.tau.ac.il/~tommer/