EMD结合Teager能量用于语音情感识别

摘要

在语音情感识别系统中,语音情感特征的提取尤为重要.在前期已有对EMD分解研究的基础上,将EMD分解与Teager能量算子相结合,用于语音情感识别.首先利用EMD分解得到一组IMF分量,再对各阶IMF分量提取Teager能量；然后通过对不同语种的不同情感语音的Teager能量在Mel频率的分析,提出了一种新的情感特征.基于EMD分解的Me1频率的Teager能量谱系数(ETMC),最后利用SVM分类方法对不同语种的不同情感进行识别.实验结果表明,该方法有很好的识别结果.%In speech emotion recognition system,the extraction of speech emotion feature is especially important,on the basis of studies of EMD,combining the EMD decomposition with Teager energy operator for speech emotion recognition.First,the EMD decomposition are used to get a set of IMF component,and for each IMF component extraction Teager energy,then through to the different emotions of different languages Teager energy in Mel frequency analysis,a new emotional characteristic is proposed:based on the EMD decomposition Teager energy coefficient (ETMC),finally using the SVM classification method to identify different emotions of different languages,experimental results show that this method has a good recognition results.