Authors: Jian-Da Wu, Yi-Jang Tsai, Chih-Wei Chuang, Li-Hung Fang and Di-En Song
This paper presents a speaker identification system using the Wigner-Ville distribution (WVD) feature extraction and artificial neural network. In earlier work, the Wigner-Ville distribution was often used to analyze the non-stationary signal, because it supplies an uncomplicated and plain energy spectrum diagram both in the time and frequency domains. These instantaneous energy diagrams presented the magnitude of each speaker under various speech signal conditions. The features were used as inputs to neural network classifiers for speaker identification. In this study, the back-propagation neural network (BPNN) and generalized regression neural network (GRNN) were applied to verify the performances and the training time in the proposed system. The experimental results indicated the GRNN can achieve better recognition rate performance with feature extraction using the WVD method than BPNN. The experimental results show the proposed speaker identification is effective and the performance is satisfactory.
Keywords: Speaker identification; Wigner-Ville distribution; back-propagation neural network; generalized regression neural network.