Get your tin foil hats, Orson Wells' '1984', Alex Jones podcast and believe everything you were told by 'Coast to Coast' radio shows.
This new system is capable of snapping 36 million images of your face in a second..yes, a motherf'n second. Then again how often are 36 million people's faces are looking the same direction? Well, it depends where Kate Upton is i suppose.
Stay tuned after the jump for a video display of the actual software in action
Step one of achieving a dystopia is having the physical infrastructure to monitor large amounts of people at all times. Step two is having the software (or manpower, I guess) to parse it all. A Japanese surveillance company has just made huge strides on that second step. The company, Hitachi Kokusai Electric, is just finishing development of a facial recognition system that, given enough footage, can scan and index around 36 million faces in just around 1 second. You’d need a pretty insane amount of video before that calculation time became non-trivial.
The scanning technology is pretty versatile. While it can’t identify people who are facing away from a camera — who could, really — when it comes to people who aren’t looking dead on, the software can handle 30 degrees of give, both horizontally and vertically, in either direction. It also requires the faces to be at least 40 by 40 pixels, but other than that, you’re good to go, and there is plenty you can do with the indexed information.
Once you have a face identified with the system, you can click the thumbnail and receive a wealth of search results, including every other clip the identified individual is in. This allows you to pick out someone sketchy in one scene and immediately see where they were before and after, plus it also allows you to upload a suitable photo and then use the system to see if the person is, or isn’t, in the area. That is, so long as they’ve glanced towards a camera.
Obviously, the system is going to be marketed towards entities with both a need for wide scale surveillance and a large amount of footage to sift through. The company has suggested it might be useful for applications in public transit, large shopping centers, and of course, law enforcement. The system isn’t available at this very moment, but it’s slated to go on the market sometime next fiscal year.
Of course, the existence of technology that can perform facial recognition on such a scale doesn’t directly translate to immediate and total surveillance of everyone everywhere. It does, on the other hand, make the installation of cameras, lots of cameras, infinitely more useful and justifiable, from a cost-benefit standpoint. There’s no word as to the cost of the system, but you can bet it won’t be cheap, so it will probably only be put to use in the largest of large communities and businesses. Even so, the fact remains that anyone who has ever thought "man, I’d love to install 1,000 more cameras, but I can’t sift through all that video" will have a potential solution very soon.