论文部分内容阅读
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.
Emotional computing is the key to emotional speech recognition.Experimental modal decomposition (EMD) algorithm is the core of Hilbert-Huang Transform (HHT) algorithm, using the piecewise power function interpolation algorithm to find the emotion envelope, can achieve better emotion recognition effect Employing software MATLAB to simulate the EMD (Emotional Mode Decomposition) feature of Emotion Speech Signal, we can get a Emotion Envelope by EMD decomposition of Emotion Speech Signal, The purpose of emotional recognition.