Scientists develop software that creates ‘studio-style’

ADS

It’s not just teenage girls who are guilty of selfie-filtering action - scientists have now got in on the trend too.

Researchers in Massachusetts have developed an algorithm that they claim takes standard filtering techniques to a new level.

The algorithm allows the distinctive style of studio photography to be layered over the top of portraits.

Scroll down for video 

Researchers in Massachusetts have developed an algorithm that allows the style of studio photography to be layered over the top of portraits.The original photos (far left) have three styles transferred onto them: (from left) high contrast, warm and soft lighting, and high-contrasted black-and-white

The problem with existing style-transfer apps is that they tend not to work well with close-ups of faces, according to YiChang Shih, an MIT graduate.

‘Most previous methods are global: From this example, you figure out some global parameters, like exposure, colour shift, global contrast,’ Mr Shih says.

‘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.’

MIT researchers show you how to take the best selfies

The technique, known as ‘local transfer’ is thought to work well with video footage. The inset images show the style that the algorithm layered over the portraits

Using off-the-shelf face recognition software, researchers first identified a portrait, in the desired style, that has characteristics similar to those of the photo to be modified.

‘We then find a dense correspondence - like eyes to eyes, beard to beard, skin to skin - and do this local transfer,’ Mr Shih explained.

The technique, known as ‘local transfer’ is also thought to work well with video footage.

Even local transfer, however, still failed to make modified photos look fully natural, Mr Shih said.

So the researchers added another feature to their algorithm, which they call ‘multi-scale matching.’

‘Human faces consist of textures of different scales,’ Mr Shih said. ‘You want the small scale - which corresponds to face pores and hairs - to be similar, but you also want the large scale to be similar - like nose, mouth, lighting.’

Even local transfer, however, still failed to make modified photos look fully natural, Mr Shih said. So the researchers added another feature to their algorithm, which they call ‘multi-scale matching'. This image shows how a face can be transfromed to look as if it has had studio lighting (centre) 

Mr Shih said that the technique works best when the source and target images are well matched — and when they’re not, the results can be bizarre

Modifying a photo at one scale can undo modifications at another. So for each new image, the algorithm generates a representation called a Laplacian pyramid.

This allows it to identify distinctive characteristics of different scales that tend to vary independently of each other. It then concentrates its modifications on those.

Mr Shih said that the technique works best when the source and target images are well matched — and when they’re not, the results can be bizarre.

For instance, the algorithm might superimpose wrinkles on a child’s face. But in experiments involving 94 photos culled from the Flickr photo-sharing site, their algorithm performed well.

Robert Bailey, now a senior innovator at Adobe’s Disruptive Innovation Group said that the researchers’ technique is an advance on conventional image filtration.

‘You can’t get stylisations that are this strong with those kinds of filters,’ he said.

‘You can increase the contrast, you can make it look grungy, but you’re not going to fundamentally be able to change the lighting effect on the face.’

By contrast, the new technique ‘can be quite dramatic,’ Mr Bailey said. ‘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.’

HOW TO CHEAT AT A SELFIE: A GUIDE TO 'DIGITAL SURGERY' APPS 

Almost everyone is aware that celebrities in magazines are usually digitally enhanced to have flawless skin and dazzling white teeth, but how about your friends’ photos on social networks?

A new breed of photo editing app is gaining popularity as people seek to digitally doctor their selfies so that they too can looks ‘perfect’ on social media sites.

Perfect 365 - A free version allows users to get a healthy-looking natural glow and try out makeup looks, while a more comprehensive paid-for package lets people to retouch multiple people in one photo.

FaceTune - A more hands-on tool that lets people blur imperfections with little brushes. Users can brighten their teeth, zap zits and even airbrush on hair.

Modiface - This app allows people to reshape their face and enhance certain features like their eyes and lips.

Pixtr - The iPhone app simply blurs and removes imperfections in photographs.

Visage Lab - A 'beauty laboratory' that removes wrinkles and spots as well as adds digital makeup including eyeliner to photos

However, as social media is meant to be a spontaneous snapshot of life, many people are not freely admitting to using technology to make themselves look better.

While countless selfies are posted on social media networks such as Twitter and Instagram every day, relatively few people disclose they have used photo touch-up technology by using a hashtag such as #Perfect365. 

+2

A new breed of photo editing app is gaining popularity as people seek to digitally doctor their selfies so that they too can looks ‘perfect’ on social media sites. These before and after shots are from the Pixtr app

Share or comment on this article

ADS

Share this article :

Share on FB Tweet Share on G+ Submit to Digg