Text Independent

The core technology that supports VoiceSense is a text independent voiceprint verification engine. Based on this unique feature, VoiceSense can be configured to work in different operation modes, such as:

• Free Speech - The user is free to say anything while the verification is taking place.
• One-Time Password - The system asks the user to repeat a random phrase - a one-time password. This mode prevents the replay attack.

Language Independent

VoiceSense is also totally independent of the language and accent that the user speaks in.

Fast Enrollment

Unlike conventional voiceprint products that often require the speaker to engage in a lengthy training process, VoiceSense virtually does not require any training. The speaker only needs to record a short message with the system, typically in the length of a few seconds.

Adaptive Learning

VoiceSense system can adapt voiceprints over time to take into account the ageing effects on an individual's voice.

Works with all types of phones

The same caller can be accurately verified regardless of the phone they are using - a mobile phone, land line, computer microphone or even the VOIP.

Key Benefits

• Accurate and reliable identity verification
• Substantial elimination of the risk of identify theft and fraud
• Flexible enrolment and verification processes
• A better customer experience
• Scalable and cost effective
• Highly scalable

Key Features

Enrollment

The enrollment of VoiceSense is simple and short. During the enrolment, the user is asked to speak freely (Free Speech mode), or to repeat a selection of keywords provided by the system (One-Time Password mode). This one-time enrollment process takes less than one minute, and a unique voiceprint is then created for the user.

Authentication

The authentication process is equally simple and brief. When the authentication is needed, the user is asked to speak freely (Free Speech mode), or to repeat a one-time password provided by the system (One-Time Password mode). A real-time voiceprint is then created and compared with the voiceprint stored in the system's voiceprint repository.