ID R&D patents new method for detecting voice spoof attacks and deepfakes
A new patent has been granted to ID R&D for a method of detecting deepfakes and spoof attacks against voice biometrics systems.
‘Method and Devices for Detecting a Spoofing Attack’ was approved by the U.S. Patent and Trademark Organization.
The method it describes the division of a voice sample into frames, and subjected to analysis of different features for real-time detection of recorded or synthetic speech during conversation. It involves applying both a Fast Fourier Transform and a Discrete cosine transform to the speech signal.
The newly patented techniques are applied in ID R&D’s IDLive Voice software.
“True to our name, research is a strategic priority at ID R&D, and our products are built upon investigation and discovery of new ways to reduce fraud without burdening users,” comments Alexey Khitrov, CEO and co-founder of ID R&D. “This most recent patent exemplifies the role of our research in being first-to-market with solutions that make access to digital services safer and easier for millions of people globally.”
The emergence of intelligent conversational chat technology raises the possibility of automated voice phishing attacks, with ‘virtual fraudsters’ stealing voice samples from unsuspecting consumers, the company says in the announcement.
ID R&D also launched software to protect face biometric systems from deepfakes and digital injection attacks earlier this year.