Speaker recognition based on a weighted acoustic discrimination
暂无分享,去创建一个
We combine multiple-mixture single-state Markov models with phonetic classification in order to improve the performance of a speaker recognition system. Three broad phonetic classes: voiced frames, unvoiced frames and transitions, are defined. We design speaker templates by the parallel connection of the weighted outputs of three single state HMM's. Each model corresponds with a distinct sound class and the output weights take into account the perceptual influences across phonetic classes. The preliminary results show that this novel architecture outperforms its counterpart without phonetic classification.
[1] R. Tucker,et al. Voice activity detection using a periodicity measure , 1992 .
[2] S. Furui,et al. Comparison of text-independent speaker recognition methods using vector-quantization distortion and discrete and continuous HMMs , 1994 .
[3] M. Savic,et al. Variable parameter speaker verification system based on hidden Markov modeling , 1990, International Conference on Acoustics, Speech, and Signal Processing.