Multi-Classifier Speech Emotion Recognition System

This article describes the design and application of a speech emotion recognition system. The system is trained in Czech emotionally coloured speech. The output of the system is the evaluation one of the four emotional states. Classification of emotion from extracted features is performed by a multi-classifier whose structure consists of three sub-classifiers fusion by Bayes Believe rule. The proposed system was deployed in the Secure Mobile Communication Infrastructure developed for the Czech Republic Security Components.