When Tommer Leyvand was a graduate student at Tel-Aviv University, he did some interesting work on software that “enhanc[es] the aesthetic appeal (or the attractiveness) of human faces in frontal photographs (portraits).” Now a part of Microsoft’s Xbox 360 XNA team in Redmond, Leyvand’s “beautification engine” has attracted attention with a write up in the Skin Deep column of The New York Times.
The story, teased on the front page of the paper today, says the beautification engine “uses a mathematical formula to alter the original form into a theoretically more attractive version, while maintaining what programmers call an ‘unmistakable similarity’ to the original.”
“The software program, developed by computer scientists in Israel, is based on the responses of 68 men and women, age 25 to 40, from Israel and Germany, who viewed photographs of white male and female faces and picked the most attractive ones.
“Scientists took the data and applied an algorithm involving 234 measurements between facial features, including the distances between lips and chin, the forehead and the eyes, or between the eyes.
“Essentially, they trained a computer to determine, for each individual face, the most attractive set of distances and then choose the ideal closest to the original face. Unlike other research with formulas for facial attractiveness, this program does not produce one ideal for a feature, say a certain eye width or chin length.
“They ran the photographs of 92 women and 33 men through the engine, creating before and after shots — essentially, a computer-generated version of hot or not.”
The Times’ coverage was supplemented by a slide show of before and after shots of movie stars and Leyvand himself.
Leyvand and colleagues first submitted their research on Digital Face Beautification as a technical sketch to SIGGRAPH, a premiere computer graphics research conference, in 2006.
A longer paper on the work was part of this year’s SIGGRAPH. It was highlighted in a press release on the event.
Check out Leyvand’s Tel-Aviv University page to see more example images, watch a video and read about his research.