VMM-3™ Voice Biometric Engine
The heart of our platform, the result of over 10 years of
research and development by VBG
From statistical models to the latest Deep Learning techniques, the VMM-3 engine has multiple algorithms and feature sets to meet any need.
We have both threshold-based and probability-based scoring options, coupled with adaptation techniques, to provide highly accurate scoring.
Flexibile Use Cases
Our multiple algorithms allow us to accept both Actively and Passively collected speech samples from your applications, with no loss of accuracy.
Our researchers have developed a state-of-the-art technique to separate speech from over 100 common noises - without altering the speech itself.Read about it
Our state-of-the-art components, thread-safe engine, and Java-based parallel processing combine for sub-second response times under any load.
We're actively working with a Tier 1 Research University to study and leverage the very latest tools and techniques for speech processing.