噪声自适应的多数据流复合子带语音识别方法
Noise Adaptive Multi-stream Hybrid Sub-band Approach for Robust Speech Recognition
-
摘要: 首先针对现有丢失数据语音识别技术中的边缘化(marginalisation)技术在特征运用上的局限，提出了一种倒谱特征分量的可靠性估计方法，将边缘化技术推广到常用的倒谱语音识别系统中; 然后利用基于全带和子带倒谱特征的边缘化识别器在不同噪声中的互补性能，提出了一种噪声自适应的多数据流复合子带语音识别方法。实验结果表明，所提识别方法可以自适应地选出全带和子带数据流中受噪声影响较小者并以之为主要依据进行识别，有效地提高了识别系统在多变噪声环境中的鲁棒性。Abstract: This paper first proposes a new method for evaluating the reliability of cepstral components and extends the marginalisation technique to cepstral recognizers. Then a noise adaptive multi-stream hybrid sub-band approach is proposed for robust speech recognition by making use of the complemental performances between full-band and sub-band cepstral marginalisation recognizers in different noises. Experimental results show that the proposed approach can turn to the less distorted data stream automatically and improve the robustness of the speech recognizer in various noisy environments effectively.
-
Mendel J M. White-noise estimators for seismic data processing in oil exploration[J]. IEEE Trans. Automatic Control, 1977, 22(5): 694706. .[2]Mendel J M. Optimal Seismic Deconvolution：an Estimation- based Approach[M]. New York：Academic Press，1983: 1.103.[3]Sun Shuli. Multi-sensor information fusion white noise filter weighted by scalars based on Kalman predictor[J].Automatica.2004, 40 (8):1447-[4]Sun Shuli，Deng Zili. Multi-sensor optimal information fusion Kalman filter[J]. Automatica, 2004, 40(6): 10171023. .[5]邓自立，高媛，李云等. 基于Kalman滤波的信息融合白噪声最优反卷积滤波器[J]. 科学技术与工程，2004，4(3)：169.171.[6]邓自立. 卡尔曼滤波与维纳滤波现代时间序列分析方法[M]. 哈尔滨：哈尔滨工业大学出版社，2000: 1.396.[7]邓自立. 自校正滤波理论及其应用现代时间序列分析方法[M]. 哈尔滨：哈尔滨工业大学出版社，2003: 1.343. -
计量
- 文章访问数: 2468
- HTML全文浏览量: 106
- PDF下载量: 614
- 被引次数: 0
下载:
