| 考虑性别差异的与文本无关说话人识别系统的实现 |
| |
| 引用本文： | 戴红霞,赵力.考虑性别差异的与文本无关说话人识别系统的实现[J].电声技术,2006(3):42-44. |
| |
| 作者姓名： | 戴红霞 赵力 |
| |
| 作者单位： | 1. 江苏信息职业技术学院,电子通信工程系,江苏,无锡,214061;东南大学,学习科学研究中心,江苏,南京,210096
2. 东南大学,学习科学研究中心,江苏,南京,210096 |
| |
| 基金项目： | 教育部科学技术基金;中国科学院资助项目 |
| |
| 摘 要： | 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,该方法能改善基于GMM的与文本无关说话人识别系统的识别率。
|
| 关 键 词： | 说话人识别 混合高斯模型 性别 |
| 文章编号： | 1002-8684(2006)03-0042-03 |
| 收稿时间： | 2005-09-26 |
| 修稿时间： | 2005年9月26日 |
Realization of a Text-independent Speaker Recognition System Considering the Sex Differences |
| |
| DAI Hong-xia,ZHAO Li.Realization of a Text-independent Speaker Recognition System Considering the Sex Differences[J].Audio Engineering,2006(3):42-44. |
| |
| Authors: | DAI Hong-xia ZHAO Li |
| |
| Affiliation: | 1. Department of Electronics Communications Engineering, Jiangsu Information Occupation Technical College,Wuxi 214061, China; 2. Research Center of Learning Science, Southeast University, Nanjing 210096, China |
| |
| Abstract: | Considering sex differences in the text-independent speaker recognition system based on GMM, a recognition system is proposed, by which the characteristic parameters of the speaker of different sex are classifies into two different subsets. During recognition processing, which subset the characteristic of speaker belongs to is first calculated, then further recognition in the subset is done. The result of the experiment indicates that this method can improve the recognition rate of the text-independent speaker recognition system based on GMM. |
| |
| Keywords: | speaker recognition Gaussian mixture model sex |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
