摘要
情感计算是情感语音识别的关键.经验模态分解(EMD)算法是Hilbert-Huang变换(HHT)的核心算法,采用分段幂函数插值算法求情感包络线,能达到更好的情感识别效果.利用软件MATLAB仿真了情感语音信号的经验模态分解(EMD)特性,把情感语音信号进行EMD分解后IMF做频谱变换,便能得到一个情感包络线,根据情感包络线的不同而达到情感识别的目的.
Emotion computing is the key to emotional speech recognition. Empirical mode decomposition (EMD)is the core of Hilbert-Huang transformation. Using the piecewise power function in emotional envelope, has a better effect in emotion identification. MATLAB is used to calculate the characteristic of emotional speech signal using empirical mode decomposition (EMD), we obtain an emotional envelope by transforming the IMF of emotional speech signal, and obtain a new method of emotion recognition using EMD.
出处
《数学的实践与认识》
CSCD
北大核心
2011年第5期123-128,共6页
Mathematics in Practice and Theory
基金
山西省自然科学基金(2010011020-1)
太原科技大学校青年基金(20103024)