Facebook Testing Recognition Tools For Identifying ‘Faceless’ People In Photos
No face? No problem! A new algorithm from Facebook may identify you based on your clothing, posture or hairdo.
From Facebook tagging to the newly launched Google Photos, face recognition is here to stay. If the general shape and features of your face are captured, you’ll likely be identified by modern technology. But what if your face isn’t made available? A new experimental algorithm from Facebook is trying to crack that code by looking at ulterior identifying elements.
According to NewScientist, Facebook is looking at unique characteristics such as hair, clothing, body shape and pose for identification outside of facial features. Of course facial recognition isn’t just applicable for social settings, but is also currently used in stores and churches to gain more information from their clientele. The new algorithm doesn’t need the face to make the ID.
A Facebook team that tested the new algorithm was able to pull in 40,000 public Flickr photos without faces showing and make a proper ID 83% of the time. Facebook’s head of artificial intelligence, Yann LeCun, stated:
“There are a lot of cues we use. People have characteristic aspects, even if you look at them from the back … For example, you can recognise Mark Zuckerberg very easily, because he always wears a gray T-shirt.”
This technology may not only help you tag for non-photogenic friends easier, but also may be a boon for the identification of shoppers & clientele that frequent physical locations that don’t always have a clean path to a photo.
For more information see NewScientist.
New on MarTech