Implementation of spectral maxima sound processing for cochlear implants by using Bark scale frequency band partition

Abstract

A new method based on Bark scale frequency-band partition is presented to improve the recognition performance of cochlear implants. Using the principles of physics, it is based on the human cochlea filtering properties. Also the mechanism of a cochlear implant and its spectral maxima sound processing (SMSP) strategy are presented. The time-frequency analyzing property of the human cochlea is considered. The theoretical determination of the Bark scale frequency-band is also given with a formula. Simulation results using Bark scale transform signal processing are discussed in detail and the results show that the new method is feasible in speech processors for cochlear implants.