Technoglitch
Core Member
A team of researchers has announced a novel means of user authentication for smartglasses and potentially other devices worn on the head. Called SkullConduct, the system was developed by Stefan Schneegass of the University of Stuttgart, Youssef Oualil of Saarland University, and Andreas Bulling of the Max Planck Institute for Informatics.
Via a wearable device, the system aims a special sound frequency into the user’s head, and then uses the device’s microphone to analyze how that frequency was conducted through the user’s skull; because each user’s skull shape is unique, the biometric pattern that emerges can be used for authentication. The researchers say that their tests, conducted using Google Glass headwear, were able to identify users with an accuracy rate of 97 percent.
There are some caveats, of course. The researchers’ testing used only 10 subjects, a rather small sample size. There is also some question as to whether mutable factors such as hair growth could alter results. Moreover, it was done in a controlled environment that did not feature background noise, though the researchers did establish that a one-second white noise signal “is sufficient to achieve high authentication accuracy”.
Researchers Announce Skull-Sound Authentication - FindBiometrics
Via a wearable device, the system aims a special sound frequency into the user’s head, and then uses the device’s microphone to analyze how that frequency was conducted through the user’s skull; because each user’s skull shape is unique, the biometric pattern that emerges can be used for authentication. The researchers say that their tests, conducted using Google Glass headwear, were able to identify users with an accuracy rate of 97 percent.There are some caveats, of course. The researchers’ testing used only 10 subjects, a rather small sample size. There is also some question as to whether mutable factors such as hair growth could alter results. Moreover, it was done in a controlled environment that did not feature background noise, though the researchers did establish that a one-second white noise signal “is sufficient to achieve high authentication accuracy”.
Researchers Announce Skull-Sound Authentication - FindBiometrics
