期刊文献+

基于PMC方法的鲁棒声学模型研究 被引量:1

Noise Robust Acoustic Model Research Based on PMC
下载PDF
导出
摘要 在噪声鲁棒语音识别研究中,使用并行模型结合(parallel model combination,PMC)方法得到的模型理论上能够接近匹配噪声环境模型的性能,故成为噪声鲁棒语音识别的重要研究方向.本文首先提出了一种基于前后向差分动态参数的特征MFCC—FWD—BWD,该特征满足PMC对特征构造矩阵可逆的要求.在此基础上,提出了一种用于PMC的新模型———并行子状态隐马尔可夫模型(parallel sub-state hidden Markov model,PSSHMM),该模型每个状态包含平行关系的子状态,且子状态间存在转移关系.实验表明,PSSHMM模型在各种噪声和SNR下取得了较好的识别效果,特别是对于非平稳噪声,其鲁棒性能非常显著. In noise robust speech recognition, for PMC(parallel model combination) method, the performance of the combined model can approach that of the model matching the noisy environment theoretically, so it is an important noise robust speech recognition research field. In this paper, a novel feature MFCC FWD BWD, which is based on forward-backward difference dynamic parameters, is presented to satisfy the requirement that the feature construction matrix is invertible for PMC. Based on this condition, a novel structure model named parallel sub-state hidden Markov model (PSSHMM) is presented for PMC and each state of this model has parallel sub-states with transitions. In experiments, PSSHMM achieves good results under each kind of noise and each level of SNR, especially for non-stationary noise, its robust performance is also excellent.
出处 《中国科学院研究生院学报》 CAS CSCD 2006年第5期660-664,共5页 Journal of the Graduate School of the Chinese Academy of Sciences
关键词 并行子状态 语音识别 噪声鲁棒 并行模型结合 parallel sub-state, speech recognition, noise robust, PMC
  • 相关文献

参考文献7

  • 1Kingsbury BED,Morgan N. Recognizing reverberant speech with RASTA-PLP. In:Proceedings of ICASSP-97, Munich Germany, 1997. 1259-1262.
  • 2Gomez R,Lee A, Saruwatari H, et al. Robust speech recognition with spectral subtraction in low SNR. In:Proceedings of ICSLP-04, Jeju Island,Korea,2004. 2077 - 2080.
  • 3Gales MJF, Young S. Robust continuous speech recognition using parallel model combination. IEEE Trans. Speech and Audio Processing, 1996,4(9) :352 - 359.
  • 4Huang JW,Shen JL, Lee LS. New approach for domain transformation and parameter combination for improved accuracy in parallel model combination(PMC) techniques. IEEE Trans. Speech and Audio Processing ,2001,9( 11 ) :842 - 855.
  • 5Wet F,Veth J, Boves L, et al. Additive background noise as a source of non-linear mismatch in the cepstral and log-energy domain. Computer Speech and Language ,2005,19( 1 ) :31 - 54.
  • 6张明新,陈国平,倪宏等.一种用于并行模型噪声鲁棒语音识别的特征构造方法.第八届全国人机通信会议,北京,2005.201~205.
  • 7Young S, Kershaw D, Odell J, et al. The HTK Book (for HTK V3.0). Cambridge University,2000.

同被引文献4

引证文献1

二级引证文献3

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