Vivoka has integrated ID R&D’s IDVoice solution into the latest version of its own Voice Development Kit (VDK). The VDK is now in its third iteration, and makes it easier for developers to build their own custom applications with voice biometrics capabilities.
To that end, IDVoice was appealing both for its accuracy and for its small computational footprint. The platform’s biometric engine comes in at less than 1MB, which makes it small enough to fit on a Neural Processing Unit (NPU) without any significant dip in performance. In terms of accuracy, ID R&D has achieved a false acceptance rate of only 0.01 percent, a figure that rivals a PIN code for security and reliability.
Vivoka’s VDK, meanwhile, offers a simple user interface that gives developers access to a slew of different voice utilities through a single software hub. Those voice technologies can be embedded in devices and applications to support a wide variety of use cases, and can be tested within the VDK to make sure that everything is running smoothly before it hits the market.
ID R&D’s platform can recognize short speech segments, and comes with anti-spoofing technology to thwart those who would try to commit fraud with synthetic or recorded voices. That technology was good enough to take the top spot in the Short-duration Speaker Verification (SdSV) Challenge earlier this year, and the top spot in the NIST’s ASVspoof Challenge back in 2019.
ID R&D can deliver strong results regardless of the text and language of the speaker, and is accurate enough to be used for authentication based on Android’s authentication guidelines. It can also help identify an individual speaker when many people are talking, at least when paired with Vivoka robust transcription and automatic speech recognition capabilities.
“Biometrics has the unique ability to bring instant personalization and authentication to voice-enabled applications and devices of all kinds,” said ID R&D CEO Alexey Khitrov. “Running voice biometrics at the edge brings never-before realized value to smart speakers, smart cameras, connected cars, robotics and more.”
“As the voice market continues to evolve, new use cases for speech technology will benefit from faster and easier ways for developers and companies to build and test their ideas,” added Vivoka CEO William Simonin.
Vivoka recently supplied Vuzix with speech recognition technology for the latter’s M-Series of Smart Glasses. ID R&D, on the other hand, provided voice recognition technology for Thales’ Trusted Digital Identity Service back in July.
(Originally posted on FindBiometrics)