Application of Cortical Processing Theory to Acoustical Analysis

Advanced Acoustic Concepts, Inc. (AAC) proposes to implement a computational model of human auditory processing based upon cortical theory and to demonstrate its utility to evaluate and improve systems for speech communication and automated recognition ofspectro-temporal patterns. Specifically, AAC believes that unusual brain representations and processing strategies are largely responsible for the remarkable sensitivity and robustness exhibited by humans and animals in detecting and recognizing sound innature. The models AAC intends to implement based upon biomimetic analytical and computational models have shown great promise in application to practical problems, e.g. classification of bird species using bird vocalizations. To further capture thesecapabilities AAC proposes to formulate higher level functional models of the auditory cortex resulting in new algorithms with enhanced stability and robustness. Improvements in error rates and detection probabilities for automatic speech recognition andacoustic identification. Speaker independent language, dialect and accent identification. New technologies for the hearing impaired, including improvements to cochlear implants. Accurate monitoring for failure prediction of artificial body parts, e.g.plastic heart valves. New and/or enhanced portable language translation devices. Novel approaches to data compression, resulting in high capacity information transmission on existing channels. Separation of noisy data into subcomponents, e.g. extractingone voice from many in a room. Computer chips that emulate human hearing. Improved diagnostic methods for machinery maintenance using acoustic emissions.