[1]王振力,裴凌波,于元斌.一种基于噪声对消与倒谱均值相减的鲁棒语音识别方法[J].智能系统学报,2008,3(06):552-556.
 WANG Zhen-li,PEI Ling-bo,YU Yuan-bin.A robust speech recognition method by combining noise cancelling and cepstral mean subtraction[J].CAAI Transactions on Intelligent Systems,2008,3(06):552-556.
点击复制

一种基于噪声对消与倒谱均值相减的鲁棒语音识别方法(/HTML)
分享到:

《智能系统学报》[ISSN:1673-4785/CN:23-1538/TP]

卷:
第3卷
期数:
2008年06期
页码:
552-556
栏目:
出版日期:
2008-12-25

文章信息/Info

Title:
A robust speech recognition method by combining noise cancelling and cepstral mean subtraction
文章编号:
1673-4785(2008)06-0552-05
作者:
王振力1 裴凌波2 于元斌1
1.南京国际关系学院博士后流动站,江苏南京210039; 2.工程兵指挥学院训练部,江苏徐州221004
Author(s):
WANG Zhen-li1 PEI Ling-bo2 YU Yuan-bin1
1.Postdoctoral Station, Nanjing University of International Relations, Nanjing 210039, China;2.Training Department, The Command Academy of Engineer Corps, Xuzhou 221004, China
关键词:
自适应噪声对消语音增强谱减法噪声鲁棒语音识别倒谱均值相减法
Keywords:
adaptive noise cancellingspeech enhancementspectral subtractionnoise robust speech recognitioncepstral mean subtraction
分类号:
TN912.34
文献标志码:
A
摘要:
提出一种基于语音增强算法的噪声鲁棒语音识别方法.在语音识别预处理阶段,通过噪声对消语音增强法来抑制噪声提高信噪比.然后对增强语音提取Mel频段倒谱特征参数,并在倒谱域应用倒谱均值相减处理来补偿增强语音中的失真成分和剩余噪声.实验结果表明,在低信噪比(-12~0dB)条件下,该方法对于数字语音识别具有较好的识别率,其性能明显优于基本的Mel频段倒谱参数识别器、传统的谱减法和噪声对消语音增强法.
Abstract:
A noise resistant speech recognition method based on a speech enhancement algorithm was implemented. First,it obtains the denoised speech, with significant SNR (signaltonoise ratio) improvement, by applying adaptive noise cancelling (ANC) to the pretreatment stage of speech recognition. Then Melfrequency cepstral coefficients(MFCC)are computed from the enhanced speech. Then cepstral mean subtraction (CMS) is used to compensate for components of distortion and the residual noise of the enhanced speech in the cepstral domain. When speech samples have a low SNR, ranging from 0 to 12dB, experimental results indicate that the proposed method performs better than a standard MFCC recognizer, conventional spectral subtraction (SS) and the ANC speech enhancement for digital speech recognition.

参考文献/References:

[1]STEVEN F B. Suppression of acoustic noise in speech using spectral subtraction [J]. IEEE Trans on Speech and Audio Processing, 1979, 27(2): 113-120.
[2]徐义芳, 张金杰, 姚开盛,等. 语音增强用于抗噪声语音识别[J]. 清华大学学报:自然科学版, 2001, 41(1): 41- 44. 
XU Yifang, ZHANG Jinjie, YAO Kaisheng, et al. Speech enhancement applied to speech recognition in noisy environments[J]. Journal of Tsinghua University: Science and Technology, 2001, 41(1): 41-44.
[3]丁   沛,曹志刚. 基于语音增强失真补偿的抗噪声语音识别技术[J]. 中文信息学报, 2004, 18(5): 64-69.
DING Pei, CAO Zhigang. Robust speech recognition based on the compensation of speech enhancement distortion [J]. Journal of Chinese Information Processing, 2004, 18(5): 64-69.
[4]王振力,张雄伟,郑    翔, 等. 一种新的子波域语音增强方法[J]. 信号处理, 2006, 22(3): 325-328. 
WANG Zhenli, ZHANG Xiongwei, ZHENG Xiang, et al. A new wavelet domain speech enhancement method [J]. Signal Processing, 2006, 22(3): 325-328.
[5]MAMMONE R J,ZHANG Xiaoyu,RAMACHANDRAN R P. Robust speaker recognition: a feature-based approach [J]. IEEE Signal Processing Magazine, 1996, 13(5): 58.
[6]DAVIS S B, MERMELSTEIN P. Comparison of parametric representations of monosyllabic word recognition in continuously spoken sentences [J]. IEEE Trans on Speech and Audio Processing, 1980, 28(4): 357-366.
[7]HERMANSKY H, MORGAN N. RASTA processing of speech [J]. IEEE Trans on Speech and Audio Processing, 1994, 2(4): 578-589.
[8]HERMANSKY H. Perceptual linear predictive (PLP) analysis of speech [J]. J Acoust Soc Am,1990, 87(4): 1738-1752.
[9]LIU F H, ACERO A, STERN R. Efficient joint compensation of speech for the effects of additive noise and linear filtering[C]//IEEE International Conference on Acoustics, Speech, and Signal Processing.San Francisco, USA, 1992(1): 257-260.
[10]王振力,张雄伟,杨吉斌, 等. 一种新的快速自适应滤波算法的研究[J]. 通信学报, 2005, 26(11): 1-6. 
WANG Zhenli, ZHANG Xiongwei, YANG Jibin, et al. Study of a new fast adaptive filtering algorithm [J]. Journal of China Institute of Communications, 2005, 26(11): 1-6.
[11]VIILDU O, BYE D, IAURILA K. A recursive feature vector normalization approach for robust speech recognition in noise [C]//Proceedings’ICASSP’98. Seattle, WA, USA: IEEE Acoustics, Speech and Signal Processing Society, 1998: 733-736.

备注/Memo

备注/Memo:
收稿日期:2008- 03- 06.
基金项目:江苏省博士后科研基金资助项目(0701008C);中国博士后科学基金资助项目(20070420561).
作者简介:
王振力,男,1977年生,工程师,博士后,主要研究方向为人工智能、多媒体信息处理等.发表学术论文20余篇,被SCI、EI、ISTP收录10余篇.
裴凌波,男,1972年生,讲师,主要研究方向为网络测量、网络性能建模和智能化信息检索等.发表学术论文20余篇.
于元斌,男,1973年生,讲师,博士后.主要研究方向为作战指挥.发表学术论文10余篇,出版专著3部.
更新日期/Last Update: 2009-04-06