[1]王洪海,刘 刚,郭 军.基于滑动倒谱的自动语言辨识[J].智能系统学报,2008,3(4):336-341.
WANG Hong-hai,L IU Gang,GUO Jun.Automatic language identif ication using shifted cepstra[J].CAAI Transactions on Intelligent Systems,2008,3(4):336-341.
点击复制
WANG Hong-hai,L IU Gang,GUO Jun.Automatic language identif ication using shifted cepstra[J].CAAI Transactions on Intelligent Systems,2008,3(4):336-341.
基于滑动倒谱的自动语言辨识
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
3
期数:
2008年第4期
页码:
336-341
栏目:
学术论文—自然语言处理与理解
出版日期:
2008-08-25
- Title:
- Automatic language identif ication using shifted cepstra
- 文章编号:
- 1673-4785 (2008) 04-0336-06
- Keywords:
- automatic language identification; shifted cep stra; shifted delta cep stra; Gaussian mixture model
- 分类号:
- TP391. 42
- 文献标志码:
- A
- 摘要:
- 滑动差分倒谱在自动语言辨识的研究中获得了广泛的应用. 但是滑动差分倒谱并没有利用语音信号的静态倒谱信息,在方言辨识中的研究表明静态倒谱比差分倒谱含有更多的特征信息. 为此,提出了滑动倒谱( SC)的概念, 并与滑动差分倒谱特征矢量进行了对比研究. 首先利用开发集的语音考察了滑动差分倒谱和滑动倒谱的控制参数在不同取值的情况下对识别性能的影响,利用爬山法确定了这2类特征矢量达到局部最优控制参数组合的路径,然后利用测试集的数据对优化后的2类特征矢量建立的模型进行了闭集辨识和开集辨识. 2种情况下的测试结果都表明滑动倒谱的性能优于滑动差分倒谱. 并且这2种参数还具有特征互补性,将它们进行决策级数据融合可以进一步提高系统的性能.
- Abstract:
- Shifted delta cep stra have been widely used in automatic language identification, but only delta cep strum information is emp loyed. Research on accent identification revealed that detailed cep strum ismore informative than delta cep strum. So shifted cep strum was p roposed and comparative studywas conducted between these two cep stra. Effects of their control parameters on recognition performance were investigated with speech data in the development set. The best paths of these two vectors to reach a locally op timal control parameter combination were determined with the hill2climbing method. Comparative tests performed with speech data both in the closed test set and open test set demonstrated that shifted cep stra is superior to shifted delta cep stra. In addition, they are mutually comp le2 mentary and data fusion at the decision level could further imp rove the performance of the system
- 参考文献/References:
-
[ 1 ]王洪海, 刘 刚, 郭 军. 自动语言辨识研究方法及发展概述[ J ]. 电脑与信息技术, 2007, 16 (2) : 37239.
WANG Honghai, L IU Gang, GUO Jun. Overview of ap2 p roaches to automatic language identification and recent de2 velopment [ J ]. Computer and Information Technology, 2007, 16 (2) : 37239.
?[ 2 ] SINGER E, TORRES C, GLEASON T P, et al. Acoustic phonetic and discriminative app roaches to automatic lan2 guage recognition[C ] / /Proc of Eurospeech. Geneva, 2003: 134521348.
?[ 3 ] TORRES2CARRASQU ILLO P A, SINGER E, KOHLER M A, et al. App roaches to language identification using Gaussian mixture models and shifted delta cep stral features [C ] / / Proc of ICSLP. Denver, USA, 2002: 89292.
[ 4 ]KOHLER M A. Language identification using shifted delta cep stra [ C ] / / Proc ofMidwest Symposium on Circuits and Systems. [ S. l. ] , 2002: 69272.
?[ 5 ] ZISSMAN M A. Comparison of four app roaches to automatic language identification of telephone speech [ J ]. IEEE Trans on Speech and Audio Processing, 1996, 4 (1) : 31244.
[ 6 ]BO Yin, EL IATHAMBY A, FANG Chen. Combining cep s2 tral and p rosodic features in language identification [ C ] / / Proc of 18 th International Conference on Pattern Recogni2 tion. Hongkong, 2006: 2542257.
[7 ] CAMPBELLl W, GLEASON T, NAVRATILl J, et al. Ad2 vanced language recognition using cep stra and phonotactics: MITLL System Performance on theN IST 2005 Language Recog2 nition Evaluation [ C ] / / Proc of Odyssey: The Speaker and Language RecognitionWorkshop. San Juan, Ruerto Rico, 2006: 128.
[ 8 ]BURGET L, MATEJKA P, CERNOCKY J. Discriminative training techniques for acoustic language identification [C ] / / Proc of ICASSP. [ S. l. ] , 2006: 2092212.
[ 9 ]MATEJKA P, BURGET L, SCHWARZ P, et al. Brno uni2 versity of technology system forN IST 2005 language recogni2 tion evaluation [ C ] / / Proc of Odyssey: The Speaker and Language RecognitionWorkshop. San Juan, Ruerto Rico , 2006: 57264.
?[ 10 ]WU Tingyao, COMPERNOLLE D V, DUCHATEAU J, et al. Spectral change rep resentation and feature selection for ac2 cent Identification tasks [ C ] / / Proc of the Workshop on Modeling for the Identification of Languages. Paris, 2004: 57261.
[ 11 ] FURU I S. Recent advances in speaker recognition [ C ] / / Proc of the First International Conference on Audio2 and Video2based Biometric Person Authentication. [ S. l. ] , 1997: 2372252.
[ 12 ]王洪海. 基于声学特征的自动语言辨识研究[D ]. 北京:北京邮电大学, 2007 .
WANG Honghai. Acoustic2based research on automatic language Iientification [D ]. Beijing: Beijing University of Posts and Telecommunications, 2007.
备注/Memo
收稿日期: 2007-06-28.
?基金项目:“十一五”国家863计划重点项目课题(2006AA010102)
作者简介:
王洪海, 男, 1970 年生, 高级工程师,主要研究方向为自动语言辨识,发表学术论文近10篇.
刘 刚,男, 1973年生,副教授,主要研究方向为语音识别、文字识别、语音合成等.
郭 军,男, 1959年生,教授,博士生导师,北京市中高级职称评审委员会计算机组副组长,主要研究方向为模式识别、网络控制与管理等. 主持开发的基于整形变换的手写汉字识别方法在对日本国家标准汉字数据库ETL9的测试中获最高识别率,在1995年全国评测中获得识别率第1名.
通信作者:王洪海. E-mail: greegrassw@sina. com.
更新日期/Last Update:
2009-05-18