Stressed speech emotion recognition using feature fusion of teager energy operator and MFCC

In this paper, a novel feature fusion of Teager Energy Operator (TEO) and Mel Frequency Cepstral Coefficients (MFCC), as Teager-MFCC (T-MFCC) feature extraction technique, is used to recognize the stressed emotions from speech signal. TEO is a non-linear feature of speech, basically designed to identify the stressed emotions. The Gaussian Mixture Model (GMM) is used for classification of these different stressed emotions with reference to neutral speech. It is found that the proposed method achieves better performance compared to the existing feature extraction technique like MFCC.