A non-hierarchical approach of speech emotion recognition based on enhanced wavelet coefficients and K-means clustering

This paper represents a non-hierarchical speech Emotion Recognition method, where the speaker-independent emotional features are extracted from the Teager energy (TE) operated wavelet coefficients of speech signal. The detail as well as approximate Wavelet coefficients enhanced by TE operation is used to determine entropy. Entropy values of TE operated detail and approximate wavelet coefficients downsize the feature dimension. The reduced feature vector thus formed is found effective for distinguishing different emotions when fed to a K-means clustering method in a non-hierarchical process. Detail simulations are carried out on EMO-DB German speech emotion database containing four class emotions, such as angry, happy, sad and neutral. Simulation results show that the proposed emotion recognition method provides better four-class emotion recognition performance through its attribute of speaker independence with lesser computation in comparison to a state-of the-art method.