Latest in Deepface

Image credit:

Facebook's face-recognition tech is almost as good at Stallone-spotting as you are

Jamie Rigg, @jmerigg
March 18, 2014
Share
Tweet
Share

Sponsored Links

Facebook's long been interested in facial recognition, as the photo tag-suggestion feature that didn't go down too well in Europe shows. The Zuck's social network also gobbled up a face-recognition outfit in 2012, but it's Facebook's AI research team that's made headway recently with technology that's almost as good as us meatsacks at identifying mugs. Known as DeepFace, the system uses a "nine-layer deep neural network" that's been taught to pick up on patterns by looking at over 4 million photos of more than 4,000 people. We're not as up-to-date with complex machine learning techniques as we should be, either, but the main reason DeepFace is so accurate is its method of "frontalization" -- or, creating a front-facing portrait from a more dynamic source image.

Like in the Sly storyboard above, the tech maps facial features, combines them with a generic 3D mask, and spits out a model that can be orientated so the subject is now perfectly posed. By doing this, it can simply compare the output with other images generated in the same way, thus silencing troublesome differences in lighting, expression, face position and the like as best it can. And, it does an extremely good job, with 97.25 percent recognition accuracy in a standardized test that humans score just over 97.5 percent in. DeepFace is just an academic endeavor currently, but it would seem like a wasted effort if Facebook didn't make use of all this work in the future -- because we're all looking forward to being auto-tagged in the background of poorly lit pictures with our best nightclub-faces on, aren't we?

All products recommended by Engadget are selected by our editorial team, independent of our parent company. Some of our stories include affiliate links. If you buy something through one of these links, we may earn an affiliate commission.
Comment
Comments
Share
Tweet
Share

Popular on Engadget

MIT tests autonomous 'Roboat' that can carry two passengers

MIT tests autonomous 'Roboat' that can carry two passengers

View
NASA will try to stow away its leaking asteroid sample tomorrow

NASA will try to stow away its leaking asteroid sample tomorrow

View
Microsoft's 'Mandalorian' Xbox controller will set you back $160

Microsoft's 'Mandalorian' Xbox controller will set you back $160

View
Samsung, Stanford make a 10,000PPI display that could lead to 'flawless' VR

Samsung, Stanford make a 10,000PPI display that could lead to 'flawless' VR

View
Researchers 3D-printed a cell-sized tugboat

Researchers 3D-printed a cell-sized tugboat

View

From around the web

Page 1Page 1ear iconeye iconFill 23text filevr