Speech Denoising Based on the Wavelet Neural Network of Improved Particle Swarm Algorithm
Wavelet neural network is the combination of the wavelet analysis and neural network,it combines good time-frequency localization properties of wavelet analysis and neural networks self-learning function.Thus wavelet neural network has strong approximation and fault tolerance,and has good convergence and robustness.This paper makes full use of the advantages of wavelet neural network,puts forward an improved particle swarm optimization algorithm,which applies to speech denoising research through collaboration and competition among the individuals to find the optimal solution.Finally,the Matlab simulation results prove that the particle swarm optimization algorithm is integrated into wavelet neural enhancement network is a kind of effective method to speech signal denoising,it can remove the noise to some extent,makes the characteristic peak points of original signal are well preserved and estimate the original signal preferably,this method obviously improves the speech enhancement effect.
NETL
NSTL
万方数据
维普
