Facebook’s DeepFace Project Now At Par With The Average Human In Identifying Faces

Facebook’s new AI research group reports a major improvement in face-processing software - DeepFace Project which boasts of human accuracy in identifying daces

facebook logo

facebook logo

From content posting to photo uploading, Facebook wants to understand each and every activity of yours. To make this happen for the 1.23 billion people accessing the network Facebook needs to make the news feed clever. To do so Facebook had launched an Advanced AI Effort in the third quarter of 2013.

The job of the research team was to use deep learning artificial intelligence, which simulates a neural network, to determine which post is genuinely important. The technology was also being built to sort a user’s photos, and it might even select the best shots.

Facebook’s plan is to unleash facial recognition technology with a new program that promises to identify the subject of an untagged image with nearly unparalleled accuracy. In fact some of you might have witnessed how Facebook already pulls out automatic suggestions for friends to be tagged on uploaded pictures.

The research team has now made a major breakthrough over the face matching software it has been working on. The research team now claims that humans who look at two faces can identify if they are the same person with 97.53 percent accuracy. They promise that the company’s new “DeepFace” program will be able to do the same with 97.25 percent accuracy.

While earlier Facebook’s algorithm to detect a face used to analyze the distance between an individual’s eyes and nose in both profile pictures and already tagged images. The method could be pretty easily foiled if a subject is simply tilting their head in a slightly different direction.

The new DeepFace program will be much more intensive 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.

DeepFace program will be using software to correct the angle of a face in an image, then comparing that to a 3D model of an average face. It then simulates what has been called a neural network to find a numerical description of the face. If there are enough similarities, Facebook will know if the faces are in fact the same.

Developed by Facebook artificial intelligence (AI) analysts Yaniv Taigman, Ming Yang, and Marc’ Aurelioa Ranzato, along with Lior Wolf, a faculty member at Tel Aviv University in Israel, the DeepFace is mostly for an academic pursuit. The research team behind it will present its findings at the Computer Vision and Pattern Recognition conference in Columbus, Ohio in June. The objective is to solicit the opinions of other qualified experts and gauge public opinion as a whole.

While DeepFace has not been exposed to the 1.23 billion Facebook users but then the question arises why Facebook purchased the Israeli startup Face.com for a price of approximately US$60 million.

At some point of time, Facebook would want to automate the process of identifying all your contacts, and determine who you’re photographed most frequently with, without the use of manual tagging.

Going further DeepFace program may have quite a few useful use cases but security concern is the grey area since Facebook will have the information and capacity to identify and track people on a broad scale.

As a Facebook user one might get a nightmare of being automatically tagged in a picture, the project might provide a grin on the face of national security surveillance agencies.