Application of an auditory model to the computer simulation of hearing impairment: preliminary results

C. Giguere, P.C. Woodland, A.J. Robinson

Abstract


A new computational model of the auditory periphery was recently reported by Giguere et al. (1993). The model is particularly attractive for hearing research in that it enables practical simulation of auditory nonlinearities and feedback mechanisms, and the study of the deterioration of these processes in the hearing impaired. The present authors present the results of an exploratory study designed to simulate the perceptual consequences associated with cochlear hearing loss. For this purpose, they combine the auditory model to a state-of-the-art recurrent neural network classifier to form an auditory-based automatic speech recognition system. Preliminary speech recognition results for normal and impaired operations of the auditory model are then compared. The ultimate goal of this work would be to guide the development of signal processing strategies for hearing aids. The idea is to find new ways of processing input speech so that, when passed through a model of impaired cochlea, the observed auditory nerve firing patterns and/or speech recognition scores would be as close as possible to those observed for unprocessed speech passed through a normal cochlea

Keywords


digital simulation; hearing; neural nets; physiological models; speech recognition; auditory model; computer simulation; hearing impairment; auditory periphery; perceptual; cochlear hearing loss; recurrent neural network classifier; speech recognition system; signal processing; impaired cochlea

Full Text:

PDF

Refbacks

  • There are currently no refbacks.