期刊文献+

高性能汉语数码串语音识别 被引量:9

High Performance Mandarin Digit String Speech Recognition
下载PDF
导出
摘要 本文给出了一个高性能汉语数码串非特定人连续语音识别系统 ,其声学模型基于Mel倒谱系数和连续HMM ,识别时采用多候选帧同步搜索算法 ,并采用了MCE算法进行训练以提高系统的区分能力 ,实验证明该系统的识别率为 94 8% (不定长数字串 )和 96 8% (定长数字串 ) .为增强系统的实用性 ,本文还研究了基于MAP算法的说话人自适应算法和基于置信度的拒识算法 .在进行自适应后 ,误识率可相对下降 40 %以上 ,在拒绝掉 5 %的正确语音时 ,系统识别率可以上升到 96 9% (不定长数字串 )和 98 7% (定长数字串 ) . A high performance mandarin digit string speaker-independent speech recognition system is given. The acoustic model is based on the Mel Frequency Cepstrum Coefficient and the continuous hidden Markov model (HMM). The multi-candidate frame synchronous search algorithm is adopted in the recognition stage with the MCE algorithm as the training approach. Experiments demonstrate that the correct recognition rate of the system is 94.8% (unknown length) and 96.8% (known length). In order to enhance the flexibility of the system, this paper also conduct research on the maximum a posteriori (MAP) based speaker adaptation and confidence measure based rejection. More than 40% recognition errors can be removed after adaptation and the recognition rate can be improved to 96.9% (unknown length) and 98.7% (known length) when 5% of the correct results are rejected.
出处 《电子学报》 EI CAS CSCD 北大核心 2001年第5期595-599,共5页 Acta Electronica Sinica
基金 国家自然科学基金! (No .699750 0 7) 国家 863项目! (No .863 30 6ZD1 3 0 4 6)
关键词 汉语数码串 语音识别 语音信号处理 Acoustic signal processing Algorithms Mathematical models
  • 相关文献

参考文献3

二级参考文献14

  • 1齐士钤 张家禄.汉语普通话辅音音长分析[J].声学学报,1982,(1):8-13.
  • 2王作英 曹洪.语音识别的改进隐含马尔可夫模型.863智能计算机系统主题学术会议[M].北京,1988..
  • 3计天颖.一种汉语连续语音识别的算法及其实现.博士学位论文[M].清华大学,1995..
  • 4Anastasakos A,ICASSP95,1995年,628页
  • 5计天颖,博士学位论文,1995年
  • 6Gu H,IEEE Trans On Signal Processing,1991年,39卷,8期,1743页
  • 7王作英,863智能计算机系统主题学术会议,1988年
  • 8齐士钤,声学学报,1982年,7卷,1期,8页
  • 9Hung W,EuroSpeech97,1997年,1519页
  • 10赵庆卫,清华大学学报,1997年,37卷,12期,87页

共引文献26

同被引文献43

  • 1关存太,陈永彬,吴伯修.HMM语音识别模型与一种修正训练算法[J].东南大学学报(自然科学版),1994,24(1):17-24. 被引量:2
  • 2姚天任.数字语音处理[M].武汉:华中理工大学出版社,1994.27-47.
  • 3俞铁城 周健来 宋岩涛.基于神经网络/隐马尔可夫模型的混合语音识别方法的研究现状[A]..第5届全国人机语音通讯学术会议论文集[C].哈尔滨,1998.18~21.
  • 4RabinerL JuangBH.Fundamentals of speech recognition[M].北京:清华大学出版社,1999.200-240.
  • 5LEVINSON S E, RABINER L R, SONDHI M M. An introduction to the application of the theory of probabilistic functions of a Markov process to automatic speech recognition [J]. The Bell System Technical Journal, 1983, 62 (4): 1 035 - 1 074.
  • 6Y Nonnandin et al.High-performance connected digit recognition using maximum mutual information estimation[J].lEEE Trans Speech and Audio Processing, 1994 ; 2 (2) : 299-311.
  • 7Katagiri S, Lee C H. A new HMM/LVQ hybrid algorithm for speech recognition [A]. Proc IEEE Globecom'90[C]. San Diego, CA, 1990. 1032-1036.
  • 8Ramesh P, Katagiri S, Lee C H. A new connected word recognition algorithm based on HMM/LVQ segmentation and LVQ classification[A]. Proc IEEE International Conference on Acoustics,Speech and Signal Processing (ICASSP'91) [C].Toronto, Ontario, Canada, 1991. 113-116.
  • 9Bahl L R, Brown P F, deSouza P V, et al. Maximum mutual information estimation[ A]. In:IEEE Proc ICASSP'86. Tokyo.
  • 10Nphraim Y, Demho A. Rabiner L. A minimum discrimination approach for Hidden Markov Modeling[A]. In : IEEE Proc ICASSP' 88. New York.

引证文献9

二级引证文献30

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