An Optimization Adaptive BWT Speech Enhancement Method

摘要

Since it is difficult to choose the threshold function and value for noisy speech signal after wavelet transform, this study proposed a bionic wavelet method of hierarchical threshold based on PSO. It firstly used bionic wavelet transform for wavelet decomposition of noisy speech signal. Then, the proposed threshold function was used for threshold processing and PSO algorithm was introduced to complete the hierarchical optimization. Finally, high frequency noise component separated by bionical wavelet transform is used as the input of the adaptive filter, to ensure complete removal of the signal relevant noise. Experimental results show that the method has a prominent effect of speech enhancement under different SNR conditions and achieves the optimal estimate of valuable signal and noisy components of the same frequency.