The FGC Speech Recognition Server (FGC/SRS) is a high performance multiprocessor (MP) that processes speech recognition requests which it receives over a network. The FGC Speech Libraries, which are an integral part of the FGC/SRS System, include core speech recognition functions as well as a number of auxiliary functions.

The FGC Continuous Speech Recognition Server for Telephone Companies is described in the FGC Application Brochure. We believe this is a unique and novel product based on the Company's years of research and development efforts. Each installation is scalable to meet the needs of a specific application, in that additional sets of FGC/MP-7 processors can be included in the configuration with a linear increase in performance and capacity.

The FGC/SRS System consists of a combination of Pentium Host Processors, two or more MP-7 parallel processors and the FGC Speech Library, a licensed software product described below.

THE FGC SPEECH LIBRARY

The FGC Speech Library consists of a set of libraries that include core speech recognition functions as well as auxiliary ones such as low-level audio capture. The FGC Speech Library is based on software licensed from Carnegie Mellon University, one of the leading research institutions in the United States. The FGC/SRS Speech Recognition System includes a customized data model which is created for each customer's system during the installation phase. This model has three components: a phonetic dictionary, a language model and an acoustic model.

Unique Features of the FGC/SRS Speech Recognition Server System

1.		The System can accommodate very large vocabularies of words and phrases because of its unique multiprocessor design.
2.		The System has been proven in a large public telephone network.
3.		The System is scalable to handle very large call volumes because of its modular design.
4.		The System is fault tolerant.