The Research Of The Application Of HMM Mode In The High Noise Circumstance

The technologies of speech recognition have very important senses for our lives. It can make our livesbecome more convenient and higher efficiency. During the long time research of speech recognition, peoplehave accumulated a lot of experience and knowledge. Nowadays the mainstream technology of speechrecognition should be the HMM mode. The identify rate of speech got by the template training andrecognition which disposed by the ways of HMM modal is the most high in the various of modals of speechrecognition. The process of speech recognition always includes the building of template training and thecontrast of the speech character.This article mainly studies the problems of speech recognition in high noise circumstance. Nowadays,although the technologies of speech recognition have developed so maturate, the identify rate of speech innoise circumstance is still low. This is mainly because the noise affect the speech recognition greatly, andno good way to dwindle its influence. So in this article, the wavlet is used to dispose the signal which havebeen got in the noise circumstance. A lot of signal have been collected by the recorder of WINDOWSsystem, they have been used to training the template of speech by the ways of HMM mode to get a set oftemplate of speech signal. Then the ways of HMM mode is used to recognize the signal which have beendisposed by wavlet. At the same time in order to show the influence of noise , the same signal which havenot been disposed by wavlet also is recognized by use of the ways of HMM mode . Finally, the articlecompares two different identify rate of speech got by use of two kinds of ways, at the same time analyzesthe reasons and put forward some improved suggestions.