ID R&D claims first place in global voice biometrics competition for ‘in the wild’ samples

ID R&D has claimed the top rank for speaker recognition with its IDVoice voice biometrics software in the VoxCeleb Speaker Recognition Challenge 2022 (VoxSRC-22), according to a company announcement.

Nearly 40 biometrics developers and research organizations from around the world participated in the workshop, which was held in collaboration with Interspeech 2022. Interspeech is the largest event in the world for digital speech technology, ID R&D says.

VoxSRC tests the accuracy of voice biometrics algorithms from speech recorded “in the wild,” rather than specifically for biometric matching. Audio samples consist of celebrity interviews and unedited conversations with background noise.

ID R&D says the VoXSRC 2022 results show that for currently available applications of speech audio, biometric security can be added without increasing friction for users with a passphrase requirement.

“We were excited to participate in VoxSRC for its promise of advancing biometric authentication that can be performed on whatever phrase is spoken, even in noisy environments, and so is completely transparent to the user,” comments Alexey Khitrov, CEO and co-founder of ID R&D. “Our number-one rankings at VoxSRC and several other recent voice biometrics challenges represent not only the best-in-class performance of our voice technology for a variety of use cases but also our ongoing commitment to providing security solutions that are truly frictionless.”

IDVoice was integrated with WhatsApp chatbots recently by Uruguay’s Simpletech, and also took first place in the SASV 2022 Challenge for biometric speaker verification and liveness detection.

ID R&D will demonstrate IDVoice, along with its facial and document liveness technologies, at Money20/20 USA in Las Vegas, October 24 to 26.

Article Topics

accuracy  |  algorithms  |  biometrics  |  ID R&D  |  speaker recognition  |  voice biometrics  |  VoxCeleb