Sensory adds features, improved accuracy to TrulySecure
Sensory has announced the addition of support for wake words, text independent speaker ID, easier enrollment, and major accuracy improvements to the updated TrulySecure Speaker Verification (TSSV) SDK 2.0.
The AI-powered modular biometric software component for speaker verification and identification provides profile switching, access security, and tracking conversations with personal assistants. The company says it uses advances in front-end noise compression, robust feature extraction, and spectral and temporal speaker modelling for extremely accurate and secure voice biometric enrollment and authentication, even in noisy environments.
“Sensory’s TrulySecure line of patented face and voice biometrics are a huge market success having been designed into leading banking applications, and mobile phones from leaders like Samsung, LG, and Motorola. We believe TrulySecure Speaker Verification 2.0 has a lot of market potential, especially with the addition of new security features, text independence, anti-spoof approaches and much more,” said Todd Mozer, CEO of Sensory.
TrulySecure also now features liveness detection, and the SDK now supports Python and includes significantly expanded documentation. It can also be used in combination with the company’s face biometrics for added convenience and security.
Support for wake words was enabled by a breakthrough in low-power wake word capability Sensory announced earlier this year.