| 基于TLS-NAP的文本无关说话人识别算法 |
| |
| 引用本文： | 何亮,杨毅,刘加. 基于TLS-NAP的文本无关说话人识别算法[J]. 模式识别与人工智能, 2012, 25(6): 916-921 |
| |
| 作者姓名： | 何亮 杨毅 刘加 |
| |
| 作者单位： | 清华大学电子工程系清华信息科学与技术国家实验室筹北京100084 |
| |
| 基金项目： | 国家自然科学基金项目(No.90920302,61005019,61105017);国家863计划项目(No.2008AA040201)资助 |
| |
| 摘 要： | 为提高文本无关说话人识别系统的识别率，提出一种基于总体最小二乘法的无用分量投影算法。利用总体最小二乘法估计的隐含变量考虑无用分量投影矩阵的扰动，并将该扰动最小化，使基于该隐含变量求得的投影矩阵能更好地刻画无用分量空间。在美国国家标准技术署于2008年公布说话人识别数据库上的实验结果验证该方法的有效性。
|
| 关 键 词： | 说话人识别 无用分量投影 总体最小二乘法 支持向量机 高斯混合模型 |
| 收稿时间： | 2011-11-02 |
TLS-NAP Algorithm for Text-Independent Speaker Recognition |
| |
| HE Liang,YANG Yi,LIU Jia. TLS-NAP Algorithm for Text-Independent Speaker Recognition[J]. Pattern Recognition and Artificial Intelligence, 2012, 25(6): 916-921 |
| |
| Authors: | HE Liang YANG Yi LIU Jia |
| |
| Affiliation: | Tsinghua National Laboratory for Information Science and Technology,
Department of Electronic Engineering,Tsinghua University,Beijing 100084 |
| |
| Abstract: | To improve the recognition accuracy rate of a text-independent speaker recognition system, a total least square-nuisance attribute projection (TLS-NAP) algorithm is proposed. The perturbation of the projection matrix is considered and its negative effect is minimized when hidden variables are estimated by the total least square algorithm. A better performance is obtained by the nuisance attribute projection space based on these variables. The effectiveness of the proposed method is demonstrated by the experimental results on NIST SRE 08 data corpus. |
| |
| Keywords: | Speaker Recognition Nuisance Attribute Projection Total Least Square Method Support Vector Machine Gaussian Mixture Models |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《模式识别与人工智能》浏览原始摘要信息 |
|
点击此处可从《模式识别与人工智能》下载全文 |
