Web Concepts

Posts Tagged ‘facial recognition’

Researchers at Rice University claim in a recent paper that they’ve created just such a system. Called Theia, their phone searching scheme works via a voluntarily downloaded app that allows their search tools to access the photo libraries on participating phones. When a user submits a query--say for a particular face or a particular background--Theia directs the apps on all participating phones to begin searching their photo libraries for those query subjects.

To keep things efficient and avoid overloading phones, the search is two-tiered. First, Theia uses a set of broader search parameters to determine, wholesale, which pics might be worth a closer look (these parameters can include meta data associated with images as well as the images themselves). Images that are deemed worthy of a closer look are then uploaded to Theia’s servers and subjected to greater scrutiny. Theia users can specify a processing budget on their phones so the system doesn’t dominate their processing power and slow other functions.

The idea here, of course, is to enable realtime searching of the physical world via image data provided by the crowd. Say a child is abducted; Theia springs into action searching phones near the place where the child was last seen for images that may have inadvertently captured the child, allowing authorities to begin combing through crowdsourced evidence long before images get posted to Flickr or Facebook. If it finds a positive match, it can zero in on a geographic area by focusing on other searchable cell phones in the same general area. In some cases, Theia might even allow for realtime or near-realtime tracking of a person or object through the physical world via the images snapped by multiple different cell phones.

And all it requires is that you grant some faceless entity access to the data on your smartphone. What objections could you possibly have?

[I Programmer]

The MORIS device attaches to the back of an iPhone, adding roughly 1.75 inches to the thickness of the smartphone. Police officers armed with the tool can take a photo of a person’s face from about five feet away, or scan his or her iris from about six inches, and wirelessly beam that data to law enforcement databases elsewhere to look for a match. It can also perform remote fingerprint matching.

Similar biometric technology has been deployed by the U.S. military in places like Iraq and Afghanistan to confirm the identities of civilians entering military safe zones and to search for known insurgents at checkpoints. But rolling it out in the streets of the U.S. has plenty of people concerned with privacy and Constitutional issues.

The technology lives in a somewhat gray area of the law. It’s generally permissible to take a photo of anyone in a public space, but when a law enforcement agent does so--and especially when he or she then cross references it against a criminal database--that could constitute a search, and therefore should require a warrant.

It’s another one of those situations where technology has simply outpaced the law ( you would think Ben Franklin of all people would’ve seen mobile facial recognition software coming). So while it would be nice to turn to legal precedent here, there simply is none.

Nonetheless, BI2 has deals with about 40 agencies nationwide to deliver about 1,000 of the devices starting in September. From a law enforcement standpoint, police officers seem to like it. It’s a technology that lets them get to the bottom of a situation quickly. Moreover, in the technology’s defense, it’s tough to use MORIS to abuse a person’s rights if an officer is not already in the process of abusing them.

In an interview with BI2’s chief executive Sean Mullin last year, he told PopSci that the responses of privacy groups and civil liberties advocates are entirely appropriate, but that he thinks the technology passes legal muster. The facial recognition technology requires a frontal facial image taken from close proximity, he says--in other words, it requires consent. Iris scans are practically impossible without the subject’s cooperation, as are fingerprint scans. Besides, the alternative when a police officer can’t confirm a suspect’s identity is generally a trip downtown to sort it out. MORIS simplifies that process.

Whether or not that’s enough to satisfy the privacy rights crowd--and the law--remains to be seen. How this kind of technology is treated by the law now will set the precedent for when the technology becomes more robust--and perhaps more long-range, more surreptitious, and potentially more “Big Brother.”

[WSJ]

The MORIS device attaches to the back of an iPhone, adding roughly 1.75 inches to the thickness of the smartphone. Police officers armed with the tool can take a photo of a person’s face from about five feet away, or scan his or her iris from about six inches, and wirelessly beam that data to law enforcement databases elsewhere to look for a match. It can also perform remote fingerprint matching.

Similar biometric technology has been deployed by the U.S. military in places like Iraq and Afghanistan to confirm the identities of civilians entering military safe zones and to search for known insurgents at checkpoints. But rolling it out in the streets of the U.S. has plenty of people concerned with privacy and Constitutional issues.

The technology lives in a somewhat gray area of the law. It’s generally permissible to take a photo of anyone in a public space, but when a law enforcement agent does so--and especially when he or she then cross references it against a criminal database--that could constitute a search, and therefore should require a warrant.

It’s another one of those situations where technology has simply outpaced the law ( you would think Ben Franklin of all people would’ve seen mobile facial recognition software coming). So while it would be nice to turn to legal precedent here, there simply is none.

Nonetheless, BI2 has deals with about 40 agencies nationwide to deliver about 1,000 of the devices starting in September. From a law enforcement standpoint, police officers seem to like it. It’s a technology that lets them get to the bottom of a situation quickly. Moreover, in the technology’s defense, it’s tough to use MORIS to abuse a person’s rights if an officer is not already in the process of abusing them.

In an interview with BI2’s chief executive Sean Mullin last year, he told PopSci that the responses of privacy groups and civil liberties advocates are entirely appropriate, but that he thinks the technology passes legal muster. The facial recognition technology requires a frontal facial image taken from close proximity, he says--in other words, it requires consent. Iris scans are practically impossible without the subject’s cooperation, as are fingerprint scans. Besides, the alternative when a police officer can’t confirm a suspect’s identity is generally a trip downtown to sort it out. MORIS simplifies that process.

Whether or not that’s enough to satisfy the privacy rights crowd--and the law--remains to be seen. How this kind of technology is treated by the law now will set the precedent for when the technology becomes more robust--and perhaps more long-range, more surreptitious, and potentially more “Big Brother.”

[WSJ]

SceneTap is a start-up that is using facial recognition tech to help you eschew the sausage fest at your usual watering holes and find the hangouts where you might actually have a chance of conversing with someone of the gender you’re looking for (whatever gender that might be). It will do so via a handy smartphone app that will map the joints where the objects of your desire are congregating at any given moment.

It works like this: bars install facial recognition cameras at their entrances and exits. These cameras, we are assured, are not equipped with good enough technology to actually identify you or cross reference images with something like Facebook. They simply detect gender. And in doing so, they keep a running tally of how many guys and dolls are in a given juke joint at a given time.

Using the accompanying app, you can get a good read on what your chances are at a particular place before you pay cab fare across town or drop a cover charge on some velvet-roped joint that turns out to be empty. But that’s as far as SceneTap will get you. As far as actually approaching the bar and tapping him/her on the shoulder--unfortunately for you, Casanova, there’s no app for that.

[Forbes]

At distances up to 50 yards, the glasses can reportedly scan 400 faces per second, comparing 46,000 biometric points on a person’s face against a database of terrorists and other criminals. If a match is made it is indicated by a red light that appears within the glasses frame, allowing police to zero in on those people with problematic pasts (or currently questionable legal statuses) without having to put police and citizens through the tedium of random ID checks.

As far as crowd security is concerned, it’s a pretty cool piece of technology if it works as advertised. And Brazil is due to have some big crowds passing through in the next few years. Aside from being an international tourist destination year round, cops in Rio will have to secure both the World Cup in 2014 and the Olympic Games in 2016, and police there hope to have the technology widely deployed by that point.

[AOL News]

For its part, Google is trying to get in front of the privacy argument that is undoubtedly coming (Google is getting pretty good at this by now) by assuring users that they will have to opt into such a service by checking a box. And the search giant is working on added layers of security and privacy to ensure that only those who want to be photographically found will be.

The idea is that Google’s massive search resources could be used to trawl social networks, online photo sharing sites like Flickr and Picasa, and the like to associate an individual’s face with his or her online presence. This, of course, could also include contact info like email addresses and phone numbers. It would at the very least identify a person by name, with which any reasonably tech-savvy person could track down contact information anyhow.

The technology, a Google engineer says, already exists and has for some time (though we really already knew that, since Google is not the first to come up with such an app). But the company is sensitive to privacy issues--if not out of genuine concern for its users then for avoiding public backlash--and wants to make sure that when it launches, it launches the right way. As such, Google has not said when it will release the product, or even so much as offered a rough production timeline. But don’t worry about being left out of the loop; when the app does launch, the stranger sitting across the coffee shop will surely email you to to let you know.

Update: A Google rep. dropped us a call to let us know that the news story upon which this blog post is based is "extremely speculative," and that the company has no such application in the pipeline, nor does it have plans to introduce such a facial recognition app to its development pipeline. According to Google, the engineer--director of engineering Hartmut Neven--was speaking to CNN's reporter about hypothetical uses for some of Google's technology, not about existing applications or applications under development.

[CNN]

The QuOD, as it's called, comes from the Department of Computer Science and Engineering at the University of Notre Dame, is essentially 3-D facial recognition software. It creates "face tracks," some kind of recognizable pattern in a face, to pick individual people out of a crowd. The interesting part is that it works from any video stream, a big bonus in a world where our first impulse at seeing a crime is to whip out our smartphones and document it on video.

That means that the QuOD is useful for picking out those who show up at crime scenes repeatedly, but also at picking out faces in areas documented with multiple security cameras or personal video recordings. The software counts the number of times a particular face pops up, and if it's more than a certain amount (set by the operating authorities), it'll present that person as a possible person of interest--though it doesn't seem to link to any database, like this FBI face detector does.

Luckily, that's all it does--there's no Minority Report-style motive assessment. It's simply a shorthand for a formerly tedious and arduous task, that of scanning through footage to try to find any recognizable faces. The QuOD hasn't as yet been picked up for use by any police force, but we could see it being a useful timesaver.

[Kurzweil]