MINERVA SYSTEMS & TECHNOLOGIES LLC — Department of Defense STTR Phase I: AF15-AT17

Description

ABSTRACT: Speech recognition technology is in wide use today and has been successfully integrated in a number of applications. Most of these applications require a microphone located near the talker. However, when a distant microphone is used where the speaker is at some distance from the microphone as in a hands-free communication, or in a meeting room, there is a major problem with the captured speech signals caused by interfering sounds and reverberations due to the large speaker to microphone distance. While traveling from the speaker to the microphone, the wavefront of the speech is repeatedly reflected at the walls or other objects in the room. These reflections, perceived as reverberation, alter the acoustic characteristics of the original speech. The recognition accuracy of reverberant speech by Automatic Speech Recognition (ASR) system degrades considerably. Minerva Systems & Technologies has teamed with Professor Zahorian of Binghamton university to develop novel approaches to improve reverberant speech recognition, which we will demonstrate using the sponsor provided reverberant speech data. In addition, we plan to demonstrate speech quality improvements by the proposed approaches. Once the problem of reverberant speech is solved successfully, it will open up a lot of opportunities for speech recognition and other applications.; BENEFIT: Improved signal intelligibility and quality for commercial and military applications, and for machine recognition of speech and speakers in a number of civil and military applications.