A Neural Network Based Adaptive Speech Enhancement System According to Estimation of Noise Intensity

In speech recognition under noisy environment, it is necessary to construct the system which reduces the noise and enhances the speech. Then it is effective to imitate the human auditory system which has an excellent analytical mechanism of spectrum for speech enhancement. In this paper, we propose an adaptive method using the auditory mechanism which is called lateral inhibition. This method first estimates the intensity of the noise intensity by neural network, then adjusts adaptively both the coefficients of the lateral inhibition and the adjusting coefficient of amplitude component according to the intensity of noise for each input frame. It is confirmed that this method is effective for the speech degraded not only by white noise but also by colored noise, judging from the spectral distortion measurement.