1:1 verification
Matching with a single voiceprint for authentication solutions.
BioVox features
See BioVox in action!
In this video you can see actually working BioVox through all the stages involved in any voice biometrics system, from user enrollment to speaker verification and identification. It shows an interactive demo recorded live using our on-line platform BioVox.cloud.
After the initial enrollment, we test the system both with the authentic user and also with an impostor speaker trying to impersonate him. We throw at the biometric engine a few surprises as well!
How it works
BioVox does voice biometrics in two successive steps.
Enrolment
The speech signal is analysed and the features that make each person’s voice unique are extracted.
Then a pattern is identified and modelled into a voiceprint, which comprises more than 1,000 parameters.
The voiceprint is stored and used in future verification operations.
Recognition
A new speech signal is processed and their biometric features extracted.
These features are then compared with a single voiceprint to resolve whether there is a match or not (1:1 verification).
Alternatively the features are compared with all the voiceprints available in the system to find the best match (1:N identification). You can also retrieve several candidates ordered by their matching scores to get a better picture of the situation.
Desktop, cloud & smartphone apps
Based on the BioVox voice biometrics core engine.
VoiceUnlock is our security app for Android, based on the BioVox Mobile speaker recognition engine. Now you can unlock your smartphone easily and securely, just with your voice.
VoiceUnlock doesn’t need an Internet connection: all the calculations are done locally in your device, from the initial voiceprint creation during enrolment to the voice unlock operations. This critical feature is possible thanks to the efficiency of the BioVox Mobile biometric engine.
VoiceUnlock is available in Free and Premium versions. The Premium version includes additional customization options and the ability to automatically improve the voiceprint each time it’s used.
BioVox.cloud is a complete speaker verification and identification solution provided as a convenient, flexible and easy to use SaaS. Developed together with our partner Cestel, it’s based on the BioVox core engine and exports an easy to integrate REST API that covers all the functionality, from the enrolment stage to the speaker authentication.
In addition to the REST API, you can test all the speaker recognition features right away thanks to a GUI based demo area in the website, both from pre-recorded audio files or live from the microphone.
BioVox.cloud also includes an admin area from where you can manage users, applications and check usage statistics.
Try it here: https://biovox.cloud