Application of nonlinear dynamics characterization to emotional speech

Abstract

This paper proposes the application of nonlinear measures based on nonlinear dynamics for emotional speech characterization. Measures such as mutual information, dimension correlation, entropy correlation, Shannon entropy, Lempel-Ziv complexity and Hurst exponent are extracted from the samples of a database of emotional speech. Then, statistics such as mean, standard deviation, skewness and kurtosis are applied on the extracted measures. Experiments were conducted on the Berlin emotional speech database for a threeclass problem (neutral, fear and anger emotional states). Feature selection is accomplished to select a reduced number of features. In order to evaluate the discrimination ability of the selected features a neural network classifier is used. A global success rate of 93.78% is obtained.