语音识别中小波分析的应用

语音信号是一种时变、非平稳的随机信号。现有语音识别系统的适应性比较差，尤其在高噪声环境下的语音识别非常困难，若将语音识别系统商品化，还有识别速度、拒识问题以及关键词／句检测技术等细节问题需要解决。如何把语言学、生理学、心理学方面的研究成果量化、建模并用于语音识别，还需深入研究。本文介绍了语音识别中小波分析的研究现状及几种常见的技术方法，并且分析探讨了语音识别中小波分析的应用和发展前景。

Application of wavelet analysis in speech recognition

The speech signal is a fime-varying, non stationary random signal.The existing voice recognition system adaptability is poor, especially in speech recognition in noise environment is very difficult, if the speech recognition system commercialization, and the recognition speed, to question sentence detection technology and other details and keywords / need to solve general problems.How to linguistics, physiology, psychology research achievements, and quantitative modeling for speech recognition, still needs further study.This paper introduces the technical method and several common situation of＇speech recognition research of wavelet analysis, and discusses the application and development prospect of speech recognition, wavelet analysis.