Facebook can recognise you from those pictures where your face is not conspicuous thanks to an experimental algorithm concocted in its artificial intelligence lab. Unique features like a person’s hair style, body shape, and clothing are put into perspective to determine the identity of the person.
The head of artificial intelligence at Facebook, Yann LeCun, explains that they wanted to see whether modern face-recognition algorithms could be improved to recognise people when their faces are not quite clear.
“There are a lot of cues we use. People have characteristic aspects, even if you look at them from the back,” LeCun says. “For example, you can recognise Mark Zuckerberg very easily, because he always wears a gray T-shirt.”
The team of researchers took around 40,000 public pictures from Flickr and ran them through a neural network. It was found that the final algorithm could identify people with 83 % accuracy. The algorithm was presented this month at the Computer Vision and Pattern Recognition conference in Boston, Massachusetts.
This algorithm might be used in photo apps – such as Facebook’s Moments – in the future. This application goes through the pictures on the user’s phone, sifts through them and categorises them into different events, and tags everyone it can identify.
LeCun also explains that the algorithm can also be used to alert privacy-conscious users when an obscured picture of them is uploaded on the Internet. On the other hand, users concerned about privacy might also argue that the tool has negative implications.
“If, even when you hide your face, you can be successfully linked to your identify, that will certainly concern people,” says Ralph Gross at Carnegie Mellon University in Pittsburgh, Pennsylvania, who says the algorithm is impressive. “Now is a time when it’s important to discuss these questions.”