The time sequence modeling ability of HMM and classifying ability of WNN were applied in the speaker recognition system independent from the text.

• 该方法利用HMM的时序建模能力以及小波神经网络较强的模式分类能力；进行与文本无关的说话人的识别.