Latest Posts

Xerox is developing technology that will make selecting a flattering image a matter of math...

Sometimes when I go to an event, someone will snap my picture, and it gets put up on a photo site for all the world to see (should anyone care to Google for it), whether I look good or not. Same with Facebook—I will untag myself from pictures from certain bad hair color years or those don't show me on my good side. But what if there were a way to ensure that EVERY photo taken of you was a good one? 

We might be in luck: Xerox is developing technology that will make selecting a flattering image a matter of math: When an photo is scanned through the Xerox software, an algorithm determines if the picture is a "good" one, based on certain metrics. For example, whether it's too blurry, poorly lit, or overly monochromatic. "What [the technology] shows is that now you don't need a human to select images that are going to be judged beautiful," says MIT associate professor Aude Olivia "You can run the algorithm, and it will give a good estimate."

But there is obviously a big difference between a good picture and good picture of you. Can this software actually ascertain whether the image in question shows you in your best light, and at your most attractive? Not exactly, says researcher Craig Saunders of the Xerox Research Center Europe in Grenoble, France. The software tends to choose photos that are the most compositionally appealing. "If you had a selection of reasonable shots and the software was asked to pick the best, this would be very difficult," he says.