一种噪声环境下的语音识别方法(线性预测误差法)的研究  

A Study on Noisy Speech Recognition (Linear Predictive Coding Prediction Error)

作　　者：冯成林[1] 吴淑珍[1] 

机构地区：[1]北京大学电子学系,北京100871

出　　处：《北京大学学报（自然科学版）》2000年第5期665-671,共7页Acta Scientiarum Naturalium Universitatis Pekinensis

摘　　要：介绍一种平稳噪声环境下语音识别的新的方法。该方法利用噪声的LPC系数去预测语音信号 ,从而得到LPC预测序列 ,然后把它代替原语音序列来进行语音端点的检测、语音特征的提取和在合适的匹配方式下的识别。实验结果表明 :该法在噪声环境下自动检测语音端点和提取语音信号的特征是可行的 ,获得了很满意的识别率。This paper presents a new method for speech recognition in stationary noise. By using the LPC coefficients of noise to predict all the speech signal, the method gets the LPC prediction error(LPCPE) sequence.Then use it to substitute the spee ch sequence to detect the speech terminal、extract the speech features and to re cognize in a suitable way. Isolated words recognition experiment based on DTW sh ows:this method is o.k. in automatic detection of speech terminal and extraction of speech feature; and achieves very satisfying recognition rate.

关 键 词：线性预测编码 线性预测误差 噪声环境 语音识别 

分 类 号：TN912.3[电子电信—通信与信息系统]