Office of Naval Research Selects Voci

October 18, 2011

Tools

Advanced Speaker Identification Technology Being Developed by Leader in Accelerated Speech Recognition

PITTSBURGH--(BUSINESS WIRE)-- Voci™ Technologies Inc. today announced that it has been selected by the Office of Naval Research to begin the research and development for an advanced speaker identification (SID) system. The US and other governments look for clues in conversations as they attempt to thwart acts of fraud or terrorism, and the massive amount of unlabeled speech presents a significant challenge to this effort. Traditional enrollment-based SID systems require speaker registration, severely limiting their application for fraud and crime detection applications. Voci’s approach will help overcome the challenge of associating speech with individuals who may want to avoid being identified. The Voci work is funded under Phase I of a multiphase Small Business Innovation Research (SBIR) award.

The accelerated processing speeds of the Voci speech recognition appliances will allow clustering, tracking and analysis of speech on a speaker-by-speaker basis within huge databases of recorded audio. Once the clustering is done, live conversations can be correlated in real time with the categorized and sorted voice files. Voci’s proprietary automated speaker clustering system technology will associate individual speakers to audio without requiring speaker registration. Targeted applications include identifying individuals based on voice for national security and business intelligence applications.

“Voci is honored to have been chosen by the Office of Naval Research to address the need for automatic and rapid discovery of individuals who might pose threats to our national security,” commented Anthony Gadient, Voci president and CEO. “The hardware acceleration foundation of Voci’s speech technology lends itself to this important program.”

Voci’s advanced SID system uses techniques based on acoustic tags that are uniquely associated with every speaker, similar to other biometric identifiers such as fingerprint and iris recognition. It enables the tagging of each speaker in a large database of conversations without the speaker having to initiate a connection between their name and their voice. Voci’s system will also enable customer relationship management (CRM) systems to authenticate identity without frustrating the customer by asking for passwords, PINs and tokens.

About Voci Technologies Inc.

Voci appliances accelerate the transformation of speech into text, enabling enterprise analytics platforms to discover insights contained in the spoken voice of the customer. The company pioneered the implementation of speech recognition algorithms in hardware to run orders of magnitude faster than alternatives, and at a fraction of the cost. Voci offers the most economical solution for converting voice into text, enabling customer-driven enterprises to uncover information in real time from 100% of all customer conversations captured through call centers. Existing text analytics platforms, agent dashboards and monitoring systems are empowered by integrating unstructured content from real-time telephone interactions with text from other sources. For information, visit www.vocitec.com.

Voci is a trademark of Voci Technologies Incorporated. All other trademarks are the property of their respective owners.