期刊文献+

基音周期标准化与嵌入式语音识别系统

Normalization of pitch period in embedded speech recognition system
下载PDF
导出
摘要 根据不同人发相同音节时,一个基音周期内的波形具有一定相似性的特点,提出一种新的基音周期标准化的语音信号预处理方法。该方法在一个嵌入式的、非特定人、孤立数字的语音识别系统中进行了验证,实验结果表明基音周期标准化能有效提高语音识别的准确率。 Based on the similarity of the sound wave in a pitch period when different speakers articulate the same syllables,a new pre-process method of speech signal,normalization of pitch period has been proposed in this paper.The new pre-process method has been applied in an embedded speech recognition system designed for speaker-independent isolated digitals recognition.The experiment proves that the use of normalization of pitch period can obviously increase the accuracy of the recognition rate.
出处 《计算机工程与应用》 CSCD 北大核心 2009年第10期11-12,44,共3页 Computer Engineering and Applications
基金 国家自然科学基金No.60873192~~
关键词 基音周期标准化 语音识别 嵌入式系统 normalization of pitch period speech recognition embedded system
  • 相关文献

参考文献6

二级参考文献15

  • 1郭继云,王守觉,苑海涛.一种基于频能比的端点检测算法[J].计算机工程与应用,2004,40(31):49-51. 被引量:7
  • 2[2]蔡莲红,黄德智,蔡锐.现代语音技术基础与应用.北京:清华大学出版社,2005.
  • 3[5]J C Tlmqua,B Mak.B Reaves,A robust algorithm for word boundary detection in the presence of noise.IEEE Tran on Speech and Audio Processing,1994,2(3);406-412.
  • 4Lawrence Rabiner, Biing-hwang Juang. Fundamentals of Speech Recognition. Prentice-Hall International, Inc., 1993.
  • 5Bernd Burchard, Ronald Romer. Oliver Fox. A Single Chip Phoneme Based HMM Speech Recognition System for Consumer Applications. IEEE Transactions on Consumer Electronics, 2000;46(3).
  • 6Rabiner L R, Wilpon J G, Soong F K. High Performance Connected Digit Recognition Using Hidden Markov Models.IEEE Trans. on ASSP, 1989; 37(8).
  • 7杨行峻 迟惠生 等.语音数字信号处理[M].北京:电子工业出版社,1995..
  • 8RabinerL JuangB-H.语音识别基本原理[M].北京:清华大学出版社,1999..
  • 9吴宗济,林茂灿.实验语音学教程[M].北京:高等教育出版社,1989.
  • 10J C Junqua. Robustness and cooperative multimodel man-machine communication applications[C].In:Proc Second Venaco Workshop and ESCA ETRW, 1991

共引文献21

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