基于支持向量机的汉语声调识别

Chinese Tone Recognition Based On Support Vector Machine

声调识别是汉语语音识别的重要组成部分,采用支持向量机对汉语语音进行声调识别实验,首先提取出语音的基频和能量参数,然后建立了适合支持向量机分类的等维声调特征向量,最后设计了6个二类支持向量机对测试语音进行分类识别,得到了理想的识别结果。

Tone recognition is an important of Chinese speech recognition.We looked at using support vector machine（SVM）for tone recognition in Chinese speech,Firstly,the fundamental frequency（F0）and energy parameters are extracted.Then established the equal dimension tone feature vectors for support vector machine.Finally,designed 6 binary-SVMs for speech classification and recognition,the recognition result is ideal.