Tuesday, November 20, 2012

Facial recognition in the news

This just in from RT: "In a single second, law enforcement agents can match a suspect against millions upon millions of profiles in vast detailed databases stored on the cloud. It’s all done using facial recognition, and in Southern California it’s already occurring.

Imagine the police taking a picture: any picture of a person, anywhere, and matching it on the spot in less than a second to a personalized profile, scanning millions upon millions of entries from within vast, intricate databases stored on the cloud.
It’s done with state of the art facial recognition technology, and in Southern California it’s already happening."

...

"Up to 4 million comparisons per second, per clustered server” — that’s how many matches a single computer wired to the FaceFirst system can consider in a single breath as images captured by cameras, cell phones and surveillance devices from as far as 100 feet away are fed into algorithms designed to pick out terrorists and persons of interest. In a single setting, an unlimited amount of cameras can record the movements of a crowd at 30-frames-per-second, pick out each and every face and then feed it into an equation that, ideally, finds the bad guys.

"I realized that with the right technology, we could have saved lives,” Joseph Rosenkrantz, president and CEO of FaceFirst, tells the Los Angeles Times. He says he dreamed up the project after the attacks of September 11, 2001 and has since invested years into perfecting it. Not yet mastered, however, is how to make sure innocent bystanders and anyone who wishes to stay anonymous is left alone as he expands an Orwellian infrastructure that allows anyone with the right credentials to comb through a crowd and learn facts and figures of any individual within the scope of a surveillance cam.

Speaking to reporters with Find Biometrics in August, Rosenkrantz said that the system is already in place in Panama, where computers there process nearly 20 million comparisons per second “using a FaceFirst matching cluster with a large number of live surveillance cameras on a scale beyond any other system ever implemented.”

Read the whole story by clicking here.

No comments: