Noisy Speech Recognition Based on Speech Enhancement

In speech recognition system, recognition rate is always influenced by noise because training and recognition models are often mismatch in noisy environments. In this paper, we present a recognition system based on speech enhancement. First, noisy speech is enhanced by a filter composed of morphology and wavelet, then enhanced signals are sent into recognition system based on hidden Markov model, and the model is trained by signals passed through morphology filter. Experiments show that this method can increase system performance in noisy environment.