On a scale from one to five, about how much would you say your average fashion magazine editorial or advertisement has been Photoshopped? That’s what a new algorithm, developed by a team of computer and digital forensics experts, has been created to find out.
Dartmouth professor Dr. Hany Farid and Eric Kee, a PhD student, have come up with a software tool that measures objectively just how much a photo has been altered. According to The New York Times, it takes original images and the fantasy resulting from retouching and “statistically measures how much the image of a person’s face and body has been altered.” The Times says those statistics are based on the way people interpret the change from raw image to magazine photo.
The algorithm is meant to mimic human perceptions. To do that, hundreds of people were recruited online to compare sets of before-and-after images and to determine the 1-to-5 scale, from minimally altered to starkly changed. The human rankings were used to train the software.
And those human rankings are an important metric, especially when you consider how much attention overly-Photoshopped photos have been getting. In Britain, France and Norway, legislators are campaigning to make it necessary to label altered photos that appear in magazines and advertising, so that people viewing them know they’re not looking at something that isn’t real. If that were the case, and images could be ranked, we’d think most magazines and advertising agencies would work hard to make their images to score 1s and 2s, instead of 5s.
But at this point, it’s all research. Farid and Kee’s findings are slated to be published in The Proceedings of the National Academy of Sciences, and we hope it won’t be too long before they’re turned into a real way to hold image retouchers accountable for just how much they change raw images.