Forget facial recognition technology — researchers say they can identify you from the veins on the backs of your hands.
Biometric recognition has become more prevalent in recent years. Facial recognition technology is used everywhere from airport check-in lines to police departments and even nightclubs, while iris, fingerprint and voice recognition is used across a variety of sectors, for security purposes.
But researchers from Australia’s University of New South Wales say that some biometric methods have “well known weaknesses.”
Fingerprints can be collected from a surface someone has touched and duplicated to create a dummy print, facial recognition technology could be bypassed using images garnered from social media, and contact lenses could be used to confound iris based mechanisms, Syed Shah, a researcher at the University of New South Wales’ School of Computer Science and Engineering, told CNN.
“Vein patterns lie underneath the skin, thus do not leave any imprint, unlike fingerprints, are not available over social media, unlike facial photographs, and cannot be obtained surreptitiously, unlike irises,” Shah told CNN in an email. “Therefore, we believe that a vein based approach will be much more difficult to bypass.”
Using an off-the-shelf depth camera — an Intel RealSense D415 Depth Camera — researchers took some 17,500 images from 35 people, where participants made a fist, and established the vein patterns of the hand.
Using artificial intelligence, researchers extracted “discriminating features” from these patterns — these, they say, could then be used to identify an individual with more than 99% accuracy from a group of 35 participants.
“Specially, the requirement of making (a) fist for vein extraction makes it difficult for an adversary to obtain vein patterns furtively,” Shah explained.
Shah told CNN that, while the idea of using veins to identify people is not new, it usually requires specialist technology — but his team’s research uses off-the-shelf 3D cameras.
The team from Australia say their study, published in IET Biometrics, shows the technique could be used for authenticating individuals on personal devices, like laptops and mobile phones.