14 DAY TRIAL // Face.com first appeared onto the scene in March with a Facebook application called Photo Finder that scanned through Facebook's photo albums to find the photos of users or their friends using a powerful facial-recognition algorithm. Photo Finder is still in private alpha but the company has decided to try its hand at something a little different, although based on the same technology, so it came up with Photo Tagger, which automatically tags photos of users when they are being uploaded.
While the applications themselves might not seem very complex they are powered by a special algorithm called the "hybrid descriptor-based funneled model,” which makes it possible for the apps to recognize human faces in a variety of poses, angles, lighting conditions and different facial expressions.
“Photo Tagger lets you choose albums (yours or your friends’) to scan for faces. Photo Tagger then batches the people it finds into groups and suggests tags for these people,” the announcement on the face.com blog reads. “By finding people for you, and batching them together Photo Tagger lets you tag your albums in a fraction of the time it would normally take.”
Users who upload a lot of photos should really find the app useful, as it allows them to easily tag a photo batch in a fraction of the time it would normally take. After they confirm that the tags are accurate the photos are uploaded to Facebook and the users tagged in them will be automatically notified.
While also in private alpha for now, Photo Tagger is expected to be launched publicly much sooner than Photo Finder, as the app puts less strain on face.com's servers. Tagger only scans the photos being uploaded at that time but Finder searches all of the photos available to the user through its extended network, which can quickly become a very large number, in the tens of thousands. So far though, it's progressing nicely, having scanned more than 1.5 billion photos – there are well over 10 billion on Facebook, making it the world's largest photo hosting service – and having identified over 1.4 million people in them.