Facebook’s DeepFace is a face verification system which basically Identify faces using photographs. Now the company claims this new project has reached a level where it comes pretty close to identifying faces like humans as it has shown an accuracy of 97.25%.
“In modern face recognition, the conventional pipeline consists of four stages: detect => align => represent => classify. We revisit both the alignment step and the representation step by employing explicit 3D face modeling in order to apply a piecewise affine transformation, and derive a face representation from a nine-layer deep neural network,” Facebook stated in a recent publication. “We trained it on the largest facial dataset to-date, an identity labeled dataset of four million facial images belonging to more than 4,000 identities, where each identity has an average of over a thousand samples.”
Read Also: Bionic Tentacles Moves and Picks Up Objects As If It’s Alive
Social networks like Facebook, Google+ already uses simple facial recognition tech to allow users easy photo tagging. What Facebook plans to do with DeepFace, that we do not know. But it will probably have something to do with the social network, that’s for sure. As Gawker points out, Facebook will definitely need a huge database of face photos to get this tech fully functional. Although, for a social network with over 1 billion of users, it shouldn’t be a problem.
(All images, trademarks shown on this post are the property of their respective owners)
Follow @nrjperera – Roshan Jerad Perera