Facebook’s facial recognition software has gotten smarter (via engadget). The social media giant’s algorithm now no longer needs your face to recognise you; it can identify you from the back of your head itself, or other patterns which might be a regular feature in your photo postures.
So how does that exactly happen ? Well, it will decipher the person based on ‘other clues,’ as stated by the head of artificial intelligence at Facebook. The other clues include hairstyle, pose, body shape, etc.
There are a lot of cues we use. People have characteristic aspects, even if you look at them from the back,
said Yann LeCun, head of Facebook’s artificial intelligence team.
For example, you can recognize Mark Zuckerberg very easily, because he always wears a gray T-shirt.
The software’s brand new addition is currently in testing mode. To test the new algorithm, the team pulled in 40,000 photos from Flickr, all of which were public, with some showing a person’s full body, including their face, and others partially obscuring the face.
After testing, the algorithm was able to detect who a person was, with an 83 percent accuracy rate. While details about this feature are pretty scarce, achieving an 83 percent success rate certainly seems impressive, especially given the fact that, ultimately it will become a tool of convenience for the user.
But then, even though Facebook is touting as an upgrade to its AI, the implications of this piece of facial recognition tech could be pretty far-reaching, and I honestly don’t see a legitimate use it might have in the future. The use, should be such, that it could ward off privacy concerns. Facebook’s Moments App, which helps people share photos by scanning people, has already hit rough waters in Europe over regulatory issues over the exact clause: facial recognition before creating a furore over privacy.