Noise suppression based on auditory-like filters for robust speech recognition

In this paper, an efficient noise suppression algorithm for speech signal is present which is based on auditory-like filters. The algorithm process in three steps: first, the noise energy spectral is estimated after corrupted speech is input into a set of auditory-like filters. A statistical estimation method based on multi band filters is proposed and compared with weighted average. The second step is to eliminate the estimated noise spectral from the observed signal by spectral subtraction. Finally, auditory-based feature is extracted from the enhanced signal and introduced into ASR system. The noise suppression algorithm is evaluated in speaker-dependent Chinese digit experiment and the experiment results show that the proposed algorithm, can improve the automatic speech recognition performance in noisy environment.