支持向量机的汉语连续语音声调识别方法  

Tone Recognition Based on Support Vector Machine in Continuous Mandarin Chinese

作　　者：傅德胜[1] 李仕强[2] 王水平[1] 

机构地区：[1]南京信息工程大学计算机与软件学院,南京210044 [2]南京信息工程大学信息与控制学院,南京210044

出　　处：《计算机科学》2010年第5期228-230,共3页Computer Science

摘　　要：声调信息在汉语语音识别中具有非常重要的意义。采用支持向量机对连续汉语连续语音进行声调识别实验,首先采用基于Teager能量算子和过零率的两级判别策略对连续语音进行浊音段提取,然后建立了适合于支持向量机分类模型的等维声调特征向量。使用6个二类SVM模型对非特定人汉语普通话的4种声调进行分类识别,与BP神经网络相比,支持向量机具有更高的识别率。Done is an essential component for word formation in Chinese languages. It plays a very important role in the transmission of information in speech communication. We looked at using support vector machines （SVMs） for automatic tone recognition in continuously spoken Mandarin. The voiced segments were detected based on Meager Energy Operation and ZCR. Compared with BP neural network, considerable improvement was achieved by adopting 6 binary-SVMs scheme in a speaker-independent Mandarin tone recognition system.

关 键 词：声调识别 基音频率 支持向量机 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]