3D Biometric Facial Recognition Comes To UK
Roland Piquepaille writes "In the UK, where the recent Queen's speech about national identity cards generated lots of -- mostly negative -- coverage, another potentially invasive technology is being tested with very few criticism. For example, several police departments are now testing a 3D biometric facial recognition software from Aurora, a company based near Northampton. The use of facial recognition 'is rapidly becoming the third forensic science alongside fingerprints and DNA,' according to a police officer who talked to BBC News for 'How your face could open doors.'" (More below.)
"The company claims its software is so sophisticated it can make the distinction between identical twins. And if the civil liberties groups continue to be neutral, this technology could also be deployed in airports or by private companies. Even banks are thinking to put cameras in their ATM machines to identify you. The good thing is that you will not have to remember your PIN. On the other hand, as with every new technology, is it safe for your privacy and is it possible to hack the system? Read more before making your decision."
3D Biometric Facial Recognition Comes To UK
In the UK, where the recent Queen's speech about national identity cards generated lots of -- mostly negative -- coverage, another potentially invasive technology is being tested with very few criticism. For example, several police departments are now testing a 3D biometric facial recognition software from Aurora, a company based near Northampton. The use of facial recognition "is rapidly becoming the third forensic science alongside fingerprints and DNA," according to a police officer who talked to BBC News for "How your face could open doors." The company claims its software is so sophisticated it can make the distinction between identical twins. And if the civil liberties groups continue to be neutral, this technology could also be deployed in airports or by private companies. Even banks are thinking to put cameras in their ATM machines to identify you. The good thing is that you will not have to remember your PIN. On the other hand, as with every new technology, is it safe for your privacy and is it possible to hack the system? Read more...
Here is the introduction from BBC News Magazine.
The ethical debate about identity cards has been reignited following the Queen's Speech, but its facial recognition technology is being used in other areas. Police are hailing it as a forensic breakthrough and a new "foolproof" 3D version could eventually become a routine procedure at cash machines or workplaces.
Once the preserve of science fiction, biometric facial recognition has now become a reality. Despite its association with the controversy of identity cards, it is predicted to become part of everyday life.
But is the technology ready?
As companies become more security conscious, the process of having our faces scanned is set to become more commonplace. And new technology which can produce this in a more accurate 3D form could accelerate this trend
A firm which has developed the 3D software, Aurora, claims it is sophisticated enough to distinguish between identical twins.
The brave BBC reporter tested the software for us.
I underwent the procedure myself and it only took a few seconds. A camera used a near-infrared light to put a virtual mesh on my face 16 times. It merged these into one unique template and calculated all the measurements of my features.
3D facial recognition software from Aurora Here is a computer screenshot showing you how thousands of points map your face and produce detailed measurements of what you look like
[image]
Now, the real questions are to know if the technology gives accurate results and if it's possible to hack the system.
The government's biometric trials for passports and identity cards have reportedly experienced a 10% error rate in face recognition. The Home Office denies this and says that in any case its trials were only testing the procedures and the public response, not the technology.
Aurora claims its software eliminates these alleged errors. Founder Hugh Carr-Archer says: "We can't say it's 100% but we've done tests and have a zero failure rate.
According to the police, the 3D technology is still too expensive to be widely deployed, but it continues to use successfully 2D images.
It works by scanning an image of a suspect's face - such as a CCTV picture taken from a crime scene or a drawing based on eye-witness accounts. This produces a 2D map of the face which marks attributes such as the distance between the eyes.
Then the computer uses an algorithm to compare the data of this face to thousands of others on a database of offenders - people who have ever been arrested or charged. Within seconds it lists the matches in order of relevance, just like a web search engine.
Of course, this technology is not approved by the justice and can't be used in courts. But it's used by the police
A good point they brought up on Question Time last week, they mentioned the "foolproofness" of ID cards, i think they were as optimistic as assuming a 1% failure rate.
They then went on to totally demolish this relatively high level of success by simply using numbers. There's about 60 million people in the UK, so that's 600,000 people that can be rejected. That sort of failure is just not acceptable for something such as a national ID car scheme.
the "distinquish identical twins" is hype and irrelevant.
As with all such systems it doesn't recognise faces but a metric derived from the face. It's entirely possible that two or more different faces can have the same metric (within the limits of the measuring process.)
So what do you do if someone matches your metric and is a terrorist? Unless you solve the false positive problem, and in a population of a billion people there are always going to be many false positives, you haven't solved face recognition.
This is not a theoretical problem. Already people have been falsely imprisoned because their DNA matches some found at a crime scene.
This quest for perfect identification is a waste of time and money.