Facial Recognition  

Facebook's new AI Group has been working on a deep learning project called DeepFace, to develop facial recognition software which maps 3D facial features allowing facial recognition from any angle.

Research at the newly organized Facebook AI Group has already yielded a system that recognizes faces almost as well as a human. Called DeepFace, the work is not yet a part of the Facebook system, but promises incredible possibilities for photographic recognition and other applications.

"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."

Related articles Digital Avatars Crossing The Uncanny Valley  Researchers Use Network of Cameras to Track People Indoors  AI System Breaks CATCHA Turing Test

The system has reached 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 (determined for the same dataset at a mean of 97.5%).

So far DeepFace remains purely a research project. Facebook has released a research paper on the project and the researchers will present the work at the IEEE Conference on Computer Vision and Pattern Recognition in June.

DeepFace uses two processes to recognize faces. First it corrects the angle of a face so that the person in the picture faces forward, using a default 3D model of an forward-looking face. This means the faces do not have to be in a fixed orientation to be recognized by the system.

Next the deep learning algorithms are applied using a simulated neural network works to create a numerical description of the reoriented face. If DeepFace comes up with similar enough descriptions from two different images, it decides they must show the same face.

The researchers have trained DeepFace 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.

This dataset has allowed for the accurate model-based alignment with the large facial database generalize remarkably well to faces in uncontrolled situations.

The performance of the final software was tested against a standard data set that researchers use to benchmark face-processing software, which has also been used to measure how humans fare at matching faces.

The study authors conclude:

Our work demonstrates that coupling a 3D model-based alignment with large capacity feedforward models can effectively learn from many examples to overcome the drawbacks and limitations of previous methods. The ability to present a marked improvement in face recognition, which is a central field of computer vision that is both heavily researched and rapidly progressing, attests to the potential of such coupling to become significant in other vision domains as well.

Neeraj Kumar, a researcher at the University of Washington who has worked on face verification and recognition, told MIT Technology Review that Facebook’s results show how finding enough data to feed into a large neural network can allow for significant improvements in machine-learning software. “I’d bet that a lot of the gain here comes from what deep learning generally provides: being able to leverage huge amounts of outside data in a much higher-capacity learning model,” he says.

With augmented reality increasingly becoming part of computer systems and hardware, facial recognition has the potential to be increasingly ubiquitous as well.  Systems like DeepFace hint at just how accurate such systems may be, and also point to a future where anonymity and privacy may be difficult or even impossible to maintain.


SOURCE  Facebook

By 33rd Square

Embed

 Computer Science  

A new algorithm that tags photos based on the relationships that people in images already have with each other has been developed at the University of Toronto. The algorithm uses the name and location of existing photo tags to build a "relationship graph," where personal connections in the images are calculated.

A new algorithm designed at the University of Toronto has the power to profoundly change the way we find photos among the billions on social media sites such as Facebook and Flickr.  This month, the United States Patent and Trademark Office have issued  patent #8,611,673 on this technology.

Developed by Parham Aarabi, a professor in The Edward S. Rogers Sr. Department of Electrical & Computer Engineering, and his former Master’s student Ron Appel, the search tool uses tag locations to quantify relationships between individuals, even those not tagged in any given photo.

"Essentially, we found that if people are standing close together or are tagged close together inside images – in one image it doesn't tell you a lot of information. But across hundreds of images that someone has on Facebook, it's a very good indicator of how close they are in real life, in a social sense," Aarabi told CTVNews.ca in a phone interview.

Related articles Machines May Learn Like Us  Teaching Computers To See By Seeing Like A Computer  Researchers Use Machine Learning and Genetic Algorithms To Find New Self-Assembling Nanomaterials

Imagine you and your mother are pictured together, building a sandcastle at the beach. You’re both tagged in the photo quite close together. In the next photo, you and your father are eating watermelon. You’re both tagged. Because of your close ‘tagging’ relationship with both your mother in the first picture and your father in the second, the algorithm can determine that a relationship exists between those two and quantify how strong it may be.

In a third photo, you fly a kite with both parents, but only your mother is tagged. Given the strength of your ‘tagging’ relationship with your parents, when you search for photos of your father the algorithm can return the untagged photo because of the very high likelihood he’s pictured.

“Two things are happening: we understand relationships, and we can search images better,” says Professor Aarabi.

The nimble algorithm, called relational social image search, achieves high reliability without using computationally intensive object- or facial-recognition software.

“If you want to search a trillion photos, normally that takes at least a trillion operations. It’s based on the number of photos you have,” says Aarabi. “Facebook has almost half a trillion photos, but a billion users—it’s almost a 500 order of magnitude difference. Our algorithm is simply based on the number of tags, not on the number of photos, which makes it more efficient to search than standard approaches.”

Work on this project began in 2005 in Professor Aarabi’s Mobile Applications Lab, Canada’s first lab space for mobile application development.

Currently the algorithm’s interface is primarily for research, but Aarabi aims to see it incorporated on the back-end of large image databases or social networks. “I envision the interface would be exactly like you use Facebook search—for users, nothing would change. They would just get better results,” says Aarabi.

While testing the algorithm, Aarabi and Appel discovered an unforeseen application: a new way to generate maps. They tagged a few photographs of buildings around the University of Toronto and ran them through the system with a bunch of untagged campus photos. “The result we got was of almost a pseudo-map of the campus from all these photos we had taken, which was very interesting,” says Aarabi.


SOURCE  University of Toronto

By 33rd Square

Subscribe to 33rd Square