利用性别判识技术的说话人识别系统的研究

Research on the Speaker Recognition System Considering the Sex Differences

本文在基于GMM的与文本无关说话人识别系统中考虑性别差异,提出了把不同性别的说话人特征参数归成不同的两个子集,识别时先判断待识别说话人特征是属于哪个性别子集后,再在相应的子集中进行识别的方法。实验结果表明该方法能够改善基于GMM的与文本无关说话人识别系统的识别率。

This paper considers the sex differences in the text-independent speaker recognition system based on GMM. It classifies the characteristic parameters of the speaker of different sex into two different subsets. When in recogni- tion proceszit first calculates which subset the characteristic of speaker belongs to and then does further recognition in the subset. The result of the experiment indicates that this method can improve the recognition rate of the text-independent speaker recognition system based on GMM.