首页 | 本学科首页   官方微博 | 高级检索  
     

说话人识别中的串行因子分析
引用本文:郭武,戴礼荣,王仁华. 说话人识别中的串行因子分析[J]. 模式识别与人工智能, 2009, 22(4)
作者姓名:郭武  戴礼荣  王仁华
作者单位:中国科学技术大学,电子工程与信息科学系,科大讯飞语音实验室,合肥,230027
基金项目:国家高技术研究发展计划(863计划)
摘    要:在基于因子分析的说话人识别中,提出串行训练载荷矩阵的方法.在载荷矩阵训练中,采用串行的方式训练得到说话人因子矩阵、对角阵(残差矩阵)和信道空间矩阵.在说话人注册中,将以上3个载荷矩阵拼接,采用联合估计的方法得到每个说话人的因子.采用这种策略可有效解决因子分析中的饱和问题.在NIST SRE 2006年核心测试数据库上等错误率能达到3.65%.

关 键 词:说话人识别  联合因子分析  本征音因子  等错误率(EER)

Serial Factor Analysis in Speaker Recognition
GUO Wu,DAI Li-Rong,WANG Ren-Hua. Serial Factor Analysis in Speaker Recognition[J]. Pattern Recognition and Artificial Intelligence, 2009, 22(4)
Authors:GUO Wu  DAI Li-Rong  WANG Ren-Hua
Affiliation:GUO Wu,DAI Li-Rong,WANG Ren-Hua(iFly Speech Lab,Department of Electronic Engineering and Information Science,University of Science and Technology of China,Hefei 230027)
Abstract:A serial loading matrix training method is proposed in the factor analysis based speaker recognition.In the loading matrix training process,the eigenvoice matrix,the diagonal matrix(residual) and the channel matrix are calculated serially.In the speaker enrollment process,the above three matrixes are assembled,and then the factors are calculated through the joint factor analysis.Thus,the saturation problem in factor analysis is solved.On the NIST SRE 2006 core test corpus,the equal error rate of the propose...
Keywords:Speaker Recognition  Joint Factor Analysis  Eigenvoice  Equal Error Rate(EER)  
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号