Voice is one of the biometric types approachd in the B4LL project to demonstrate the strengths of using voice and speech recognition, and therefore, the identification and authentication of mobile users to access mobile services.
The voice solution consists of an IVR Server System that allows the user to talk to a voicebot through a menu of different possibilities. The services available are a simulation of real services and have the purpose of demonstrate the use of voice registration and authentication.
Based on the B4LL project architecture, the voice solution has three main elements (call center, Engine and Biometric Supplier) and works like the figure shown below. the user calls the call center to access the menu and the available use cases. The call center communicates with the engine that acts as an orchestrator between the call center and the biometric supplier, managing the communication between them. The biometric supplier in turn provides the solution for voice recording and voice recognition.
The registration will take place through the voice recording for a specific passphrase, the voice patterns detection and speech styles detection. Thus the biometric solution creates a digital signature for the user. Once the record is created, it is possible to perform authentication using the registered passphrase by comparing the current user's voice with the stored signature.
The solution is an open source asset that allows mobile operators to test the use of voice for authenticating customers’ identity in different use cases. Specifically, the asset will allow operators and FSPs to:
Register their voice,
Navigate menus and execute commands using their voice,
Receive phone calls to validate actions/activities, and
Identify themselves using their voice to be granted permission to execute secure actions.
Figure: The figure below shows a more detailed view comparing the voice asset with our architecture.
It is important to point out that the proposed demo asset will only be available to mobile operators to see the possibilities of biometrics with demo data i.e. no real customer data will be used in this project.