[1]赵力,刘怡龙,邹采荣,等.基于VQ-HMM的无教师说话人自适应方法[J].东南大学学报(自然科学版),2001,31(2):23-26.[doi:10.3969/j.issn.1001-0505.2001.02.006]
 Zhao Li,Liu Yilong,Zou Cairong,et al.An Unsupervised Speaker Adaptation Method Based on VQ-HMM[J].Journal of Southeast University (Natural Science Edition),2001,31(2):23-26.[doi:10.3969/j.issn.1001-0505.2001.02.006]
点击复制

基于VQ-HMM的无教师说话人自适应方法()
分享到:

《东南大学学报(自然科学版)》[ISSN:1001-0505/CN:32-1178/N]

卷:
31
期数:
2001年第2期
页码:
23-26
栏目:
计算机科学与工程
出版日期:
2001-03-20

文章信息/Info

Title:
An Unsupervised Speaker Adaptation Method Based on VQ-HMM
作者:
赵力 刘怡龙 邹采荣 高西奇 吴镇扬
东南大学无线电工程系, 南京 210096
Author(s):
Zhao Li Liu Yilong Zou Cairong Gao Xiqi Wu Zhenyang
Department of Radio Engineering, Southeast University, Nanjing 210096)
关键词:
语音识别 VQ HMM 无教师说话人自适应
Keywords:
speech recognition VQ HMM unsupervised speaker adaptation
分类号:
TP391.42;TN912.3
DOI:
10.3969/j.issn.1001-0505.2001.02.006
摘要:
提出了一种新的语音识别方法,该方法综合了VQ,HMM和无教师说话人自适应算法的优点.该方法首先在每个状态通过用矢量量化误差值取代传统HMM的输出概率值来建立VQ-HMM,同时采用无教师自适应矢量量化算法,来改变VQ-HMM的各状态的码字,从而实现对未知说话人的码本适应.本文通过非特定人汉语数码(孤立和连续数码)识别实验,把新的组合方法同基于CHMM的自适应和识别方法进行了比较,实验结果表明该方法鲁棒性好,所需计算量较少,自适应和识别效果远优于基于CHMM的方法.
Abstract:
We propose a new speech recognition method by the integration of the VQ, HMM and an unsupervised speaker adaptation algorithm, it complies a VQ-distortion measure at each state instead of a discrete output probability used by a discrete HMM, and uses an adaptive VQ algorithm to alter the codewords for speaker adaptation. In this paper, the new combined method is compared with CHMM by the task of speaker-independent Chinese spoken digit (isolated/connected) recognition, and the experiments illustrate that this new method is simple and robust, and has good performance.

参考文献/References:

[1] Shore J E,Burton D K.Discrete utterance speech recognition without time alignment.IEEE Trans IT,29(4):473~491
[2] Li Zhao,Suzuki H,Nakagawa S.A comparison study of probability functions in HMMs through spoken digit recognition.IEICE TRANS INF and SYST,1995,E78-D(6):669~675
[3] Hirata Y,Nakagawa S.Speaker adaptation of continuous parameter HMM.In:IEICE,ed.International Conference Spoken Language Processing’90.Kobe,Japan:IEICE,1990.67~70
[4] Gauvian J L,Lee C H.Maximum a posteriori estimation for multivariate gaussian mixture observations of markov chains.IEEE Trans on Speech and Audio Processing,1994,2(2):291~298
[5] Lee C H,Lin C H,Silverman H F.A study on speaker adaptation of the parameters of continuous density hidden Markov models.IEEE Trans Signal Processing,1995,39(5):806~814
[6] Huo Qiang,Lee Chin-Hui.On-line adaptive learning of the continuous density hidden markov model based on approximate recursive bayes estimate.IEEE Trans on Speech and Audio Processing,1997,5(2):161~172
[7] Nakagawa S.Speech recognition by hidden Markov model using segmental statistics.IEICE Trans,1996,J79-D-II(12):2032~2038

相似文献/References:

[1]孙暐,吴镇扬.基于独立感知理论的鲁棒语音识别算法[J].东南大学学报(自然科学版),2005,35(4):506.[doi:10.3969/j.issn.1001-0505.2005.04.002]
 Sun Wei,Wu Zhenyang.Robust speech recognition algorithm based on fletcher-allen principle[J].Journal of Southeast University (Natural Science Edition),2005,35(2):506.[doi:10.3969/j.issn.1001-0505.2005.04.002]
[2]吕勇,吴镇扬.基于隐马尔可夫模型与并行模型组合的特征补偿算法[J].东南大学学报(自然科学版),2009,39(5):889.[doi:10.3969/j.issn.1001-0505.2009.05.004]
 Lü Yong,Wu Zhenyang.Feature compensation algorithm based on hidden Markov model and parallel model combination[J].Journal of Southeast University (Natural Science Edition),2009,39(2):889.[doi:10.3969/j.issn.1001-0505.2009.05.004]

备注/Memo

备注/Memo:
作者简介:赵力,男,1958年生,副教授.
更新日期/Last Update: 2001-03-20