Zehu's products are designed as software development kits (SDKs) that provide a package of APIs, libraries and tools to OEM applications.

ZEHU AUTHENTICATOR ™

Zehu Authenticator ™ is a biometric verification platform that uses a voice biometric technology for real time verification of a person's identity. By matching the user's voice to a mathematical voice model stored in a database, Zehu Authenticator ™ returns a highly accurate authentication within seconds. This is performed using one of the following three possible authentication methods:

• Fixed Sentence: a predefined sentence used to enroll and authenticate users
• Flexible Sentence: a user-selected phrase for registration and authentication
• Free Speech: users speak freely during registration and authentication

Architecture

Zehu Authenticator ™ consists of a voice verification server, a database that stores mathematical voice models, system setting and configuration, and a web-based management application, as shown in the following system architecture.



The system can be easily integrated with many applications such as IVR platforms, time and attendance systems, smart card technologies, as well as other biometric platforms.

Key Benefits

• Multiple methods of authentication
• Improved verification accuracy and reliability through analysis of independent speech factors along with feedback mechanisms such as changes in the speaker's voice due to illness, age and time of day
• Automated, quick, and inexpensive speaker verification using practical and accessible media types, such as land line telephone, cellular and VoIP phones
• Adapts to the changes in the users voice over time
• Highly flexible
• Secure against playback
• Easy to use - works virtually anywhere, without necessity for special hardware
• Natural security against hackers
• Optional two-factor authentication in a single action, such as spoken content and intonation
• Intuitive and user-friendly
• Cost effective
• Configurable verification strategies such as:
  • One-step verification
  • One-step verification with playback protection (random challenge)
  • Two-step verification
  • Text-independent verification

Main Features

• Noise filter - enables reliable authentication in noisy or non-ideal communication channels by using Digital Signal Processing technology to eliminate unwanted noise or non-speech artifacts
• Fixed sentence enrollment and authentication:
  • For enrollment, users are requested to repeat a specific sentence 3 times
  • For authentication, users are requested to repeat a specific sentence
• Flexible sentence enrollment and authentication, enabling users to select a specific sentence
  • For enrollment, users are requested to repeat a sentence of their choice 3 times
  • For authentication, users are requested to repeat the enrollment sentence
• Free speech registration and authentication (text independent)
  • For enrollment, users are requested to speak freely for about 1 minute
  • For authentication, users are requested to speak freely for about 8 seconds
• Digital playback protection which prevents the use of pre-recorded sentences
• Enhanced Security features
  • Communication between the Zehu Authenticator server and the system database can be encrypted
  • Communication between the client-application and Zehu Authenticator server can be encrypted using SSL
  • Administration activities can be assigned to different users based on their rolls and permissions
  • Secure login required when performing any system activity
• Web-based administration enabling management from virtually anywhere
• Microsoft directory services integration (Active Directory)
• Customizable, rich and easy-to-run reporting system
• Scalability and high availability by utilizing Microsoft Network Load Balancing (NLB) technology
• Intel's® Multi-Core® support for enhanced performance