Speech feature extraction of cochlear implants on the basis of auditory perception wavelet transform

A method, which is on the basis of auditory perception wavelet transform, is proposed to model the speech process and extract features for cochlear implants. First, the original speech signal is decomposed by using an auditory perception wavelet transform. Second, a linear predictive coding method is used to extract the fundamental frequency and formant frequency in the perception channel. Experimental results show that this method, compared with the traditional filter-bank methods, not only easily adjusts the parameters of the common filter-bank, but also agrees with the human hearing characteristics. At the same time, it settles pitch-formant frequency interaction and raises the time-frequency resolution and the accuracy of estimation of pitch frequency and formant.