Led by the Intelligence Advanced Research Projects Activity (IARPA) and the US National Institute of Standards and Technology (NIST), this competition is legit!
The Face Recognition Prize Challenge, the biggest face-recognition competition to date, involves two tasks: face verification and face search. The first one may be more familiar to us as phone manufacturers are trying to master this. You know, unlocking something with your face and having a software to decided for that. The latter, however, is next level problem. Every image of a person has to be found in a pile (millions) of images in a database.
For the face-verification task, a company called Ntech made it to the top. Their FindFace software has a 99.9 per cent accuracy of matching a person's face! (Now Apple probably knows what company to call for their next model.) For the face-search task, Shaghai start-up Yitu Tech won. With just one guess, its software was able to choose the right face from a million mugshots 80 per cent of the time! This is very helpful to policemen when tracking suspects in CCTV footages.
“The public need to know what the technology is capable of,” says Shuang Wu at Yitu Tech. “It’s just going to keep growing and growing. You can now find a needle in a haystack very reliably.”
All these are made possible by deep learning, "an artificial intelligence technique that uses neural networks – software loosely based on the structure of the brain – that have many different layers", of which both the winners utilized.
Get weekly science updates in your inbox!