Neglect Face ID! Subsequent-gen biometrics will take heed to the sound of your bones

etri south korea sound biometrics figure 1ETRI

From established biometric identification strategies like fingerprints or facial recognition to weirder, extra experimental ones like customers’ distinctive “heartprint” and even the form of their butts, there are many excessive tech methods know-how can be utilized to verify that an individual is who they are saying they’re. Now researchers from South Korea have added another to the combo — and it’s a doozy.


In a brand new research revealed within the journal IEEE Transactions on Cybernetics, researchers from the Electronics and Telecommunications Analysis Institute (ETRI) describe how sound waves passing by way of an individual’s physique can be utilized to supply a constructive ID. Utilizing bioacoustic modeling, it’s potential to determine somebody primarily based on metrics just like the stiffness of their joints, their pores and skin, or the load of their bones. That’s one thing that’s fairly darn arduous to faux.

“We are able to consider our physique as a musical instrument that has a novel form and composition of supplies,” Joo Yong, one of many ETRI researchers on the challenge, advised Digital Developments. “Our know-how evaluates these traits of our physique by vibrating a sure physique half — for instance, a hand — and listen to[ing] the propagating sound as we alter the frequency of excitation. Our system is subtle sufficient to extract the options of our physique as a proxy of a consumer’s anatomical and biomaterial properties and differentiate the person with excessive accuracy.”


etri south korea sound biometrics figure 1 4ETRI

The know-how is at present being examined on greater than 50 topics. Whereas it’s nonetheless too early for conclusive large-scale outcomes, Yong mentioned that the efficiency of the method to this point is much like that of state-of-the-art fingerprint and iris biometrics.

“In present biometrics — resembling fingerprint, iris, and face recognition — one could make faux copies for spoofing as a result of they depend on the structural options of the acquired picture, and subsequently as soon as a template is stolen it may be a everlasting risk to spoofing,” Yong mentioned. “Our technique makes use of traits contained in the physique and extracts data within the frequency area, which creates a brand new stage of safety and makes it helpful for functions that require a excessive stage of safety.”

As well as, Yong mentioned that the method may very well be helpful for offering seamless identification. Fairly than asking the consumer to position a physique half on or in entrance of a biometric sensor, this method might perform a steady evaluation of customers. As Web of Issues gadgets proceed to proliferate, providing us customized companies, this may occasionally show to a very helpful facet of the know-how.

Editors’ Suggestions