Instagram devotees and champions of the selfie rejoice, for your filtering options may be about to get a little more diverse. Researchers from MIT have developed an algorithm that takes standard "style-transfer" techniques up a notch, enabling the distinctive style of studio photography to be layered over the top of your portraits.

Existing approaches to style transfer (the adjusting of metrics like color temperature and contrast to change the look of an image) are described as global methods. According to the graphics researchers, these are not so effective when applied to close-up shots of human faces.

"Most previous methods are global," says YiChang Shih, an MIT graduate student. "From this example, you figure out some global parameters, like exposure, color shift, global contrast. We started with those filters but just found that they didn’t work well with human faces. Our eyes are so sensitive to human faces. We’re just intolerant to any minor errors."

With the revelation that these existing methods don't particularly lend themselves to the nuances of our freckles, dimples and and beards, Shih and his team set about devising a new method: one he describes as a "local transfer."

The team used off-the-shelf facial recognition software to find common features between the subject's face and the face in an example photograph, the characteristics of which they would try to replicate.

"We find a dense correspondence – like eyes to eyes, beard to beard, skin to skin – and do this local transfer" explains Shih.

The team experienced some complications when the example and the target images were not properly matched, such as wrinkles superimposed onto a young person's face. The researchers say, however, that after testing the algorithm over 94 photographs, it demonstrated a significant advancement on existing image filters, giving the original images the look of a professionally lit photograph.

"You can take a photo that has relatively flat lighting and bring out portrait-style pro lighting on it and remap the highlights as well," says Robert Bailey, senior innovator at Adobe's Disruptive Innovation Group, and a member of Shih's research team.

The researchers are now looking at how the technology can be put to use in consumer applications and will present their findings at Vancouver's Siggraph computer graphics conference in August.

Source: MIT