首页 | 本学科首页   官方微博 | 高级检索  
     

多通道子空间算法在说话人识别中的应用
引用本文:关海欣,曾庆宁.多通道子空间算法在说话人识别中的应用[J].声学技术,2008,27(3):396-402.
作者姓名:关海欣  曾庆宁
作者单位:桂林电子科技大学信息与通信学院,桂林,541004
摘    要:深入研究了基于多通道信号子空间的语音增强算法原理,对算法中各个参数对性能的影响进行了深入剖析.同时给出一种选取噪声方差的简单且有效的方法,并通过研究分析,证明多通道信号子空间算法不仅消噪明显而且对语音的损伤微小,而且相比于单通道子空间语音增强算法除了性能上的提升外,还没有导致计算量的增加。最后将多通道子空间语音增强算法用于说话人识别系统.并与其它多通道语音增强算法(延迟求和波束形成、波束形成后维纳滤波、线性约束最小方差波束形成)进行了对比.实验表明多通道信号子空间语音增强算法在多种噪声环境下均可有效的提高说话人识别系统的识别性能。

关 键 词:多通道信号子空间  语音增强  说话人识别
收稿时间:2006/6/29 0:00:00
修稿时间:2006/9/19 0:00:00

Application of muti-channal subspace algorithm to speaker recognition
GUAN Hai-xin and ZENG Qing-ning.Application of muti-channal subspace algorithm to speaker recognition[J].Technical Acoustics,2008,27(3):396-402.
Authors:GUAN Hai-xin and ZENG Qing-ning
Affiliation:(Guilin University of Electronic Technology, Guilin 541004, China)
Abstract:The multi-channel signal subspace algorithm for speech enhancement is studied, and the algorithm's parameters which affects the performance are analyzed. A simple and an effective method for estimating the noise variance are given. Through analysis, it is proved that the multi-channel signal subspace algorithm can denoise obviously and make very little voice injury. Compared with the single-channel subspace speech enhancement algorithm, the multi-channel signal subspace algorithm can not only upgrade the performance, but also have no increase in the amount of calculation. The muti-channel signal subspace algorithm has been used in speaker recognition system, and compared with other speech enhancement algorithms, such as Delay and Sum Beamforming, Beamforming and Wiener Filter, and Linear Constraint Minimum Variance Beamforming. The experimental results show that the muti-chanel signal subspace algorithm can effectively improve the recognition rate of the speaker recognition system in noisy environment.
Keywords:muti-channel signal subspace  speech enhancement  speaker recognition
本文献已被 维普 万方数据 等数据库收录!
点击此处可从《声学技术》浏览原始摘要信息
点击此处可从《声学技术》下载全文
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