论文部分内容阅读
提出了一种基于相对自相关序列(Relative Autocorrelation Sequences,RAS)MFCC(Mel-Frequency Ceps-tral Coefficient)特征的丢失数据带噪语音识别新方法。首先分析了环境噪声对RAS-MFCC的影响,提出了一种基于掩盖原理的不可靠分量检测方法;然后采用丢失数据(Missing data,MD)技术来消除畸变分量对识别过程的影响,实验结果表明,本文所提的识别方法可以在不同类型和信噪比的噪声环境中有效提高RAS-MFCC的识别率,并且其性能优于典型的基于滤波器组(Filter bank)语音特征的丢失数据语音识别方法。
A new speech recognition method based on Relative Autocorrelation Sequences (RAS) and MFCC (Mel-Frequency Ceps-tral Coefficient) is proposed. First of all, the influence of environmental noise on RAS-MFCC is analyzed. An unreliable component detection method based on the mask principle is proposed. Missing data (MD) technique is then used to eliminate the effect of distortion component on the recognition process. The experimental results show that , The recognition method proposed in this paper can effectively improve the recognition rate of RAS-MFCC in the noise environment of different types and signal-to-noise ratio, and its performance is better than the typical loss data speech recognition based on the speech characteristics of Filter bank method.