期刊文献+

基于矢量泰勒级数的模型自适应算法 被引量:2

Model Adaptation Algorithm Using Vector Taylor Series
下载PDF
导出
摘要 在实际环境中,由于测试环境与训练环境的不匹配,语音识别系统的性能会急剧恶化。模型自适应算法是减小环境失配影响的有效方法之一,它通过测试环境下的少量自适应数据,将HMM模型的参数变换到测试环境下。该文将矢量泰勒级数用于模型自适应,同时对HMM模型的均值向量和协方差矩阵进行变换,使其与实际环境相匹配。实验证明,该文算法优于MLLR算法和基于矢量泰勒级数的特征补偿算法,在低信噪比环境中性能提高尤为明显。 In actual environments the performance of speech recognition system may be degraded significantly because of the mismatch between the training and testing conditions. Model adaptation is an efficient approach that could reduce this mismatch, which adapts model parameters to new conditions by some adaptation data. In this paper, a new model adaptation using vector Taylor series is presented, which adapts the mean vector and covariance matrix of hidden Markov model. The experimental results show that the proposed algorithm is more effective them MLLR and the feature compensation algorithm based on vector Taylor series in various environments, especially in low signal-to-noise ratio environments.
作者 吕勇 吴镇扬
出处 《电子与信息学报》 EI CSCD 北大核心 2010年第1期107-111,共5页 Journal of Electronics & Information Technology
基金 国家973计划项目(2002CB312102) 国家自然科学基金(60971098)资助课题
关键词 语音识别 模型自适应 矢量泰勒级数 隐马尔可夫模型 Speech recognition Model adaptation Vector Taylor series Hidden Markov model
  • 相关文献

参考文献10

  • 1Moreno P J, Raj B, and Stern R M. A vector Taylor series approach for environment- independent speech recognition[C] Proc. IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Atlanta, Georgia, USA, 7-10 May 1996: 733-736. 被引量:1
  • 2Moreno P J. Speech recognition in noisy environments[D]. [Ph.D. dissertation], Carnegie Mellon University, 1996. 被引量:1
  • 3Sasou A, Asano F, and Nakamura S, et al.. HMM-based noise-robust feature compensation[J]. Speech Communication, 2006, 48(9): 1100-1111. 被引量:1
  • 4Kim W and Hansen J H L. Feature compensation in the cepstral domain employing model combination[J]. Speech Communication, 2009, 51(2): 83-96. 被引量:1
  • 5Gauvain J L and Lee C H. Maximum a posteriori estimation for multivariate Gaussian mixture observations of Markov chains[J]. IEEE Transactions on Speech and Audio Processing 1994, 2(2): 291-298. 被引量:1
  • 6Leggetter C J and Woodland P C. Maximum likelihood linear regression for speaker adaptation of continuous density hidden Markov models[J]. Computer Speech and Language, 1995, 9(2): 171-185. 被引量:1
  • 7Gales M J F and Woodland P C. Mean and variance adaptation within the MLLR framework[J]. Computer Speech and Language, 1996, 10(4): 249-264. 被引量:1
  • 8Gales M J F and Young S J. Robust speech recognition in additive and convolutional noise using parallel model combination[J]. Computer Speech and Language, 1995, 9(4): 289-307. 被引量:1
  • 9Kim D and Yook D. Linear spectral transformation for robust speech recognition using maximum mutual information[J]. IEEE Signal Processing Letters, 2007, 14(7): 496-499. 被引量:1
  • 10Li J, Deng L, and Yu D, et al.. High-performance HMM adaptation with joint compensation of additive and convolutive distortions via vector Taylor series[C]. Proc.IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU), Kyoto, Japan, 9 -13 December 2007: 65-70. 被引量:1

同被引文献25

  • 1张明新,倪宏,张东滨,陈国平.基于PMC方法的鲁棒声学模型研究[J].中国科学院研究生院学报,2006,23(5):660-664. 被引量:1
  • 2Dautrich B, Rabiner L, Martin T. On the effects of varying filter bank parameters on isolated word recognition [ J ]. Acoustics, Speech and Signal Processing, IEEE Transactions on, 1983,31 (4) :793-807. 被引量:1
  • 3Lockwood, P A. Experiments with a nonlinear spectral subtract and hidden Markov models and the projection for robust speech recognition in cars [ J ]. Speech Communication, 1992,11 (2/3) :215-228. 被引量:1
  • 4Das S, Bakis R, N6das A, et al. Influence of background noise and microphone on the performance of the IBM TANGORA speech recognition system[ C ]. Acoustics, Speech, and Signal Processing, 1993 ICASSP-93,1993 IEEE International Conference on IEEE, 1993,2:71-74. 被引量:1
  • 5Preuss R. A frequency domain noise cancelling preprocessor for narrowband speech communications systems [ C ]. Acoustics, Speech, and Signal Processing, IEEE International Conference on ICASSP"/9 IEEE, 1979,4:212-215. 被引量:1
  • 6Berouti M, Schwartz R, Makhoul J. Enhancement of speech corrupted by acoustic noise [ C 1. Acoustics, Speech, and Signal Processing, IEEE International Conference on ICASSP'79IEEE, 1979,4:208-211. 被引量:1
  • 7Agarwal A, Cheng Y M. Two-stage mel-warped Wiener filter for robust speech recognition[ C ]. Proc ASRU, 1999, 99:67-70. 被引量:1
  • 8Lira J, Oppenheim A. All-pole modeling of degraded speech[ J]. Acoustics, Speech and Signal Processing, IEEE Transactions on, 1978,26 ( 3 ) : 197-210. 被引量:1
  • 9Musicus B, Lira J. Maximum likelihood parameter estimation of noisy data [ C ]. Acoustics, Speech, and Signal Processing, IEEE International Conference on ICASSP79 IEEE, 1979,4:224-227. 被引量:1
  • 10Hansen J H, Clements M A. Constrained iterative speech enhancement with application to speech recognition [ J ]. Signal Processing, IEEE Transactions on, 1991,39 (4) :795-805. 被引量:1

引证文献2

二级引证文献7

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