Stefan Mileschin

This whole facial recognition "approaching human-level performance" is cool and all but your nine-layer deep neural network is starting to creep me out a bit there Facebook. The learned representations coupling the accurate model-based alignment with the large facial database generalize remarkably well to faces in unconstrained environments, even with a simple classifier. Our method reaches an accuracy of 97.25% on the Labeled Faces in the Wild (LFW) dataset, reducing the error of the current state of the art by more than 25%, closely approaching human-level performance.

http://www.hardocp.com/news/2014/03/...l_performance/