期刊文献+
任意字段
题名或关键词
题名
关键词
文摘
作者
第一作者
机构
刊名
分类号
参考文献
作者简介
基金资助
栏目信息

基于混合因子分析的隐马尔可夫模型 被引量：1

A Hidden Markov Model Based on Mixture of Factor Analysis
下载PDF
导出
摘要 经典隐马尔可夫模型用于语音识别存在的两个主要缺陷是“离散状态假设”和“独立分布假设”。前者忽略了语音信号的非平稳性,后者忽略了语音信号的相关性。文章将混合因子分析方法用于语音建模,提出了基于混合因子分析的隐马尔可夫模型框架,并用动态贝叶斯网络形象地表示。该模型框架不仅从理论上解决了上述问题,而且给出许多语音建模的选择。目前广泛使用的统计声学模型均可视为该模型的特例。 The “discrete states assumption”and “conditional independent assumption” are two main limitations in standard hidden Markov model for speech recognition.The former ignores time-short stationarity of speech signals and the latter disconsiders the intra-frame correlation between the feature vector elements.This paper investigates the combined traditional hidden Markov modeling technology with mixture of factor analysis.It proposes a hidden Markov model based on mixture of factor analysis（HMM-MFA） and use a dynamic Bayesian networks for this approach.A number of standard models including HMM-DG（Hidden Markov Model Based Diagonal Gaussian distributions）,HMM-FA（Hidden Markov Model Based Factor Analysis） and other currently very popular acoustic models are forms of HMM-MFA with different configurations.
作者 王新民 姚天任
机构地区 孝感学院物理系 华中科技大学电信系
出处 《计算机工程与应用》 CSCD 北大核心 2005年第24期50-52,共3页 Computer Engineering and Applications
基金 湖北省教育厅重点项目基金资助(编号:2002A02004)
关键词 隐马尔可夫模型 混合因子分析 动态贝叶斯网络 hidden Markov model,mixture of factor analysis,dynamic Bayesian networks
分类号 TN912 [电子电信—通信与信息系统]
  • 相关文献

参考文献12

  • 1Rabiner L R.A tutorial on hidden Markov models and selected application in speech recognition[J].Proceeding of IEEE, 1989;77(2) : 257-285.
  • 2Wilpon J G,Rabiner L R,Lee C H et al.Automatic recognition of keyword in unconstrained speech using hidden Markov models[J].lEEE Trans Acoust ,Speech ,Signal Processing, 1990;38( 11 ) : 1870-1878.
  • 3Rabiner L R,Juang B H.Fundamentals of Speech Recognition[M]. Englewood Cliffs, NJ : Prentice-Hall, 1993.
  • 4Mak B,Bocchieri E.Direct training of subspace distribution clustering hidden Markov model[J].IEEE Transactions on Speech and Audio Processing, 2001 ; 9 (4) : 378-387.
  • 5王新民,姚天任.基于因子分析的隐马尔可夫模型及其训练算法[J].计算机工程与应用,2004,40(15):79-81. 被引量：3
  • 6王新民,姚天任.量化子空间分布隐马尔可夫模型[J].华中科技大学学报（自然科学版）,2003,31(7):16-18. 被引量：8
  • 7王新民,黄新堂,姚天任.基于多相关分组的HMM训练算法[J].华中师范大学学报（自然科学版）,2003,37(2):179-182. 被引量：5
  • 8Bocchieri E,Mark B.Subspace distribution clustering hidden Markov model [J].IEEE Transactions on Speech and Audio Processing,2001; 9(3) :264-275.
  • 9Bellegarda J R, Nahamoo D.Tied mixture continuous parameter modeling for speech recognition[J].IEEE Trans Acoustics,Speech, Signal Processing, 1990;38(12) :2033-2206.
  • 10Gales M J F.Semi-tied covariance matrices for hidden Markov models[J].IEEE Transactions on Speech and Audio Processing,1999; 7(3):272-281.

二级参考文献22

  • 1Bocchieri E, Mark B. Subspace distribution clustering hidden Markov mold. IEEE Trans. Speech and Audio Processing, 2001, 9(3) : 264--275.
  • 2Tadahashi S, Sagayama S. Four-level tied-structure for efficient representation of acoustic modeling. Proc.IEEE Int. Conf. Acoustics, Speech, Signal processing,1995, 1:520--523.
  • 3Rabiner L, Juang B H. Fundamentals of Speech Recognition. Englewood Cliffs, NJ: Prentice-Hall, 1993.
  • 4Bellegarda J R, Nahamoo D. Tied mixture continuous parameter modeling for speech recognition. IEEE Trans. Acoustics, Speech,Signal Processing, 1990, 38:2033-2045.
  • 5Takahashi S, Sagayama S. Effects of variance tying for four-level tied structure phone models. Proc. ASI Conf., 1995, 1:141--142.
  • 6Huang X, Jack M A. Semi-continuous Hidden Markov models for speech signals. Journal of computer speech and language, 1989, 3(3): 239--251.
  • 7Xiaolin L, Parizeau M, Plamon R. Training hidden Markov models with multiple observations--A combinatorial method[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000,22(4 ) : 371 - 377.
  • 8Baggenstoss P M. A modified Baum-Welch algorithm for hidden Markov models with multiple observation spaces [J].IEEE Transactions on Speech and Audio Processing, 2001,9(4) :411-416.
  • 9Bocchieri E, Mark B. Subspace distribution clustering hidden Markov model [J]. IEEE Transactions on Speech and Audio Processing,2001,9(3) :264-275.
  • 10Rabiner L, Juang B H. Fundamentals of Speech Recognition[M]. Indianapolis : Prentice-Hall, 1993.

共引文献9

同被引文献13

  • 1王新民,姚天任.基于因子分析的隐马尔可夫模型及其训练算法[J].计算机工程与应用,2004,40(15):79-81. 被引量：3
  • 2L E Baum, T Petrie. A maximization technique occurring in the statistical analysis of probabilistic functions of Markov chains [J]. Annals of Math. Statistic (S0003-4851), 1970, 41(1): 164-171.
  • 3L E Baum. An inequality and associated maximization technique in statistical estimation for probabilistic functions of Markov processes [J]. Inequalities (S0003-4851), 1972, 3(1): 1-8.
  • 4A P Dempster, N M Laird, D B Rubin. Maximum likelihood from incomplete data via EM algorithm [J]. Journal of Royal Statistical Society B (S1548-1352), 1977, 39(1): 1-38.
  • 5C F J Wu. On the convergence properties of the EM algorithm [J]. Annals of statistics (S0090-5364), 1983, 11(1): 95-103.
  • 6S E Levinson, L R Rabiner, M M Sondhi. An Introduction to the Application of the Theory of Probabilistic Function of Markov Process to Automatic Speech Recognition [J]. Bell System Technical (S0005-8580), 1983, 62(4): 1035-1074.
  • 7Wilpon J G, Rabiner L R, Lee C H, Goldman E. Automatic recognition of keyword in unconstrained speech using hidden Markov models [J]. IEEE Trans. Acousticm, Speech, Signal Processing (S0096-3518), 1990, 38(11): 1870-1878.
  • 8L R Rabiner, B H Juang. Fundamentals of Speech Recognition [M]. Englewood Cliffs, NJ: Prentice-Hall, 1993.
  • 9B Mak, E Bocchieri. Direct training of subspace distribution clustering hidden Markov model [J]. IEEE Transactions on Speech and Audio Processingm (S1063-6676), 2001, 9(4): 378-387.
  • 10E Bocchieri, B Mark. Subspace distribution clustering hidden Markov model [J]. IEEE Transactions on Speech and Audio Processing (S1063-6676), 2001, 9(3): 264-275.

引证文献1

计算机工程与应用
2005年 第24期

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部