FBI launches facial recognition database

I wonder how well this is going to play with social media photo tagging, because you know they will try to integrate it.

I can’t wait until someone gets arrested because their Facebook photo looks like a criminal’s photo.

(via NextGov)

The FBI by mid-January will activate a nationwide facial recognition service in select states that will allow local police to identify unknown subjects in photos, bureau officials told Nextgov.

The federal government is embarking on a multiyear, $1 billion dollar overhaul of the FBI’s existing fingerprint database to more quickly and accurately identify suspects, partly through applying other biometric markers, such as iris scans and voice recordings.

Often law enforcement authorities will “have a photo of a person and for whatever reason they just don’t know who it is [but they know] this is clearly the missing link to our case,” said Nick Megna, a unit chief at the FBI’s criminal justice information services division. The new facial recognition service can help provide that missing link by retrieving a list of mug shots ranked in order of similarity to the features of the subject in the photo.

Today, an agent would have to already know the name of an individual to pull up the suspect’s mug shot from among the 10 million shots stored in the bureau’s existing Integrated Automated Fingerprint Identification System. Using the new Next-Generation Identification system that is under development, law enforcement analysts will be able to upload a photo of an unknown person; choose a desired number of results from two to 50 mug shots; and, within 15 minutes, receive identified mugs to inspect for potential matches. Users typically will request 20 candidates, Megna said. The service does not provide a direct match.

Michigan, Washington, Florida and North Carolina will participate in a test of the new search tool this winter before it is offered to criminal justice professionals across the country in 2014 as part of NGI. The project, which was awarded to Lockheed Martin Corp. in 2008, already has upgraded the FBI’s fingerprint matching service.

Local authorities have the choice to file mug shots with the FBI as part of the booking process. The bureau expects its collection of shots to rival its repository of 70 million fingerprints once more officers are aware of the facial search’s capabilities.

Thomas E. Bush III, who helped develop NGI’s system requirements when he served as assistant director of the CJIS division between 2005 and 2009, said, “The idea was to be able to plug and play with these identifiers and biometrics.” Law enforcement personnel saw value in facial recognition and the technology was maturing, said the 33-year FBI veteran who now serves as a private consultant.

NGI’s incremental construction seems to align with the White House’s push to deploy new information technology in phases so features can be scrapped if they don’t meet expectations or run over budget.

But immigrant rights groups have raised concerns that the Homeland Security Department, which exchanges digital prints with the FBI, will abuse the new facial recognition component. Currently, a controversial DHS immigrant fingerprinting program called Secure Communities runs FBI prints from booked offenders against the department’s IDENT biometric database to check whether they are in the country illegally. Homeland Security officials say they extradite only the most dangerous aliens, including convicted murderers and rapists. But critics say the FBI-DHS print swapping ensnares as many foreigners as possible, including those whose charges are minor or are ultimately dismissed.

Megna said Homeland Security is not part of the facial recognition pilot. But, Bush said in the future NGI’s data, including the photos, will be accessible by Homeland Security’s IDENT.

The planned addition of facial searches worries Sunita Patel, a staff attorney with the Center for Constitutional Rights, who said, “Any database of personal identity information is bound to have mistakes. And with the most personal immutable traits like our facial features and fingerprints, the public can’t afford a mistake.”

In addition, Patel said she is concerned about the involvement of local police in information sharing for federal immigration enforcement purposes. “The federal government is using local cops to create a massive surveillance system,” she said.

Bush said, “We do have the capability to search against each other’s systems,” but added, “if you don’t come to the attention of law enforcement you don’t have anything to fear from these systems.”

Other civil liberties advocates questioned whether the facial recognition application would retrieve mug shots of those who have simply been arrested. “It might be appropriate to have nonconvicted people out of that system,” said Jim Harper, director of information policy at the libertarian Cato Institute. FBI officials declined to comment on the recommendation.

Harper also noted large-scale searches may generate a lot of false positives, or incorrect matches. Facial recognition “is more accurate with a Google or a Facebook, because they will have anywhere from a half-dozen to a dozen pictures of an individual, whereas I imagine the FBI has one or two mug shots,” he said.

FBI officials would not disclose the name of the search product or the vendor, but said they gained insights on the technique’s accuracy by studying research from the National Institute of Standards and Technology.
In responding to concerns about the creation of a Big Brother database for tracking innocent Americans, Megna said the system will not alter the FBI’s authorities or the way it conducts business. “This doesn’t change or create any new exchanges of data,” he said. “It only provides [law enforcement] with a new service to determine what photos are of interest to them.”

In 2008, the FBI released a privacy impact assessment summarizing its appraisal of controls in place to ensure compliance with federal privacy regulations. Megna said that, during meetings with the CJIS Advisory Policy Board and the National Crime Prevention and Privacy Compact Council, “we haven’t gotten a whole lot of pushback on the photo capability.”

The FBI has an elaborate system of checks and balances to guard fingerprints, palm prints, mug shots and all manner of criminal history data, he said.

“This is not something where we want to collect a bunch of surveillance film” and enter it in the system, Megna said. “That would be useless to us. It would be useless to our users.”

Facebook has a new role

Questions about what social networks mean for personal privacy and security have been brought to a head by research at Carnegie Mellon University that shows that Facebook has essentially become a worldwide photo identification database.  Paired with related research, we’re looking at the prospect where good, bad and ugly actors will be able identify a face in a crowd and know sensitive personal information about that person.

Read the whole article on Forbes.

Google Group Members to Use Facial Recognition to Identify London Rioters

(TechCrunch)

A new Google Group called “London Riots Facial Recognition” has appeared online, in the wake of the riots that rocked the U.K. capital over the weekend. The group’s goal is to use facial recognition technologies to identify the looters who appear in online photos.

The group appears to be thoughtfully considering its actions, in threads titled “Ethical Issues,” and “Keeping Things Legal,” for example. They’ve also stated that “it’s important we only use legal sources for images.”

However, there’s a major “creepy” factor to this undertaking, too. The idea that a group of people would team up online to use (misuse?) facial recognition technologies in this way, notably outside professional law enforcement channels, seems like a modern take on vigilante style justice, where the torches of the angry villagers have turned into APIs and algorithms.

In one newer thread, started just this morning, a commenter offers their assistance in building a tool using the Face.API, which could help identify people in photos posted on Facebook, Flickr and Twitter. There is even talk of using the Facebook Graph API and the Twitter API in conjunction with the Face.com one to help better identify the criminals.

While clearly, we have nothing against criminals being brought to justice, there still may be some concerns involved with this type of online behavior. As argued here on Hacker News, this method could incriminate people who were not participating, but were bystanders, or simply trying to get home. Whether their actions here are legal, whether or not they involve public photos, the question is – do we want to crowdsource justice in this way?

Real-time facial recognition pulls multiple identity aspects

This week at the Black Hat security conference researchers from Carnegie Mellon University demonstrated how facial recognition technology can be used to positively identify a person and possibly even to gain access to their personal information.

The Carnegie team used three relatively simple technologies to create their face recognition system: An off-the-shelf face recognizer, cloud computing processing, and personal data available through the public feed at social networking sites such as Facebook.

They were even able to reproduce some of the effects using a smartphone app that overlayed parts of their deduced ID data onto a view of the world in real time using an augmented reality app.

Read more at Fast Company.

One click to a 3-D face

(via Wired)

Don’t bother with the iris scanner or the fingerprinting machine. Leave the satellite-enabled locators and tell-tale scents back on the base, military manhunters. If an Air Force plan works out as planned, all you’ll need to track your prey is a single camera, snapping a few seconds of footage from far, far away.

Huntsville, Alabama’s Photon-X, Inc. recently received an Air Force contract to develop such a camera. With one snap, the company claims, its sensor can build a three-dimensional image of a person’s face: the cornerstone of a distinctive “bio-signature” that can be used to track that person anywhere. With a few frames more, the device can capture that face’s unique facial muscle motions, and turn those movements into a “behaviormetric” profile that’s even more accurate.

“The proposed work will help identify non-cooperative dismounts using remote sensors, from standoff distances that were previously impossible,” reports Toyon Research Corporation, which also got an Air Force grant for bio-signature development. “This identity information can help intelligence analysts connect specific people to events and locations, and learn about insurgent operations.”

But the combo won’t just help flesh-and-blood airmen keep tabs on their fellow humans, Photon-X adds. It “can help Humanoid Robots navigate and find objects in a cluttered room.”

And it could be used to monitor suspicious behavior practically anywhere. “A brief list of potential industries includes law enforcement, banking, private corporations, schools and universities, casinos, theme parks, retail, and hospitality.”

So the next time you’re at the Bellagio, and a real-life android walks up to you, calls you by name, and asks about what you thought was an extremely private evening — you can thank the Air Force and its tech-enabled manhunters for the experience.