That's a drastic departure from
existing voice biometric mechanisms, which require training from each individual who will use them, said Kassem Fawaz, who worked on the project as a graduate student at U-M and is now an assistant professor at the University of Wisconsin.