28th Febraury 2006

Biometric Security Announces VoiceVault Version 5.0

First widely-available version of ISO 17799-accredited voice verification solution delivers remote two-factor authentication over any telephone

LONDON, UK – Tuesday, 28th February 2006 - Biometric Security Limited, the leader in voice-based biometric verification security and identity management, today announced the general availability of version 5.0 of its VoiceVault™ voice verification platform.  VoiceVault is the only voice verification platform to be accredited to the ISO 17799 standard.

VoiceVault Version 5.0 is the result of continued investment in science, mathematics and technical development over the past 12 months. It is the first release since VoiceVault was acquired by Biometric Security Limited in 2005 and reflects the company’s commitment to extending it s technological lead and delivering market leading biometric voice verification solutions to customers.

Vance Harris, Chief Technology Officer responsible for VoiceVault, said: “The new services and enhancements delivered in this release of VoiceVault offers our customers products and managed services that are at the forefront of identity verification, management and enhanced, easy-to-use security.  Voice offers significant benefits as a cost-effective and reliable means of biometric verification . Services that do not require the user to be present, such as ‘card not present’ credit/debit card transactions, have been crying out for this kind of solution.”

Version 5.0 introduces the following new services and enhancements:

• Automatic Number and Pass-phrase Verification:  Enhancing the challenge response verification options available within the product and solutions, Version 5.0 introduces Automatic Number Verification (verification from a spoken number) and a pass-phrase verification (using, for example, everyday spoken words or place names ), allowing for remote, two-factor authentication.

• Refinement of intra-speaker variance: Existing versions of VoiceVault’s biometric engine have been capable of identifying variances in a speaker’s voice that occur as a result of ordinary ailments, such as the common cold.  Version 5.0 can now identify more intra-speaker variances  (for example differences in voice that occur naturally during different times of the day, after drinking, or smoking) enabling the engine to model a speaker’s voiceprint more consistently, which in turn may lead to a reduction in the number of calls having to be resolved manually.