Getting Bored with HTK? Using HMMs for Emotion Recognition from Speech Signals

During the past years the classification and recognition of emotions has evolved to a significant topic of interest within the research on natural spoken dialogue systems. In this article we pick up the idea of using hidden Markov models (HMMs) to recognize emotions from speech signals taken from the database of German emotional speech. At first we give an overlook on different characteristics of selected emotions with respect to the features extracted from the speech signal. Then we describe the experimental setup and we present current results of different scenarios which include the separation of female and male speakers or different thresholds in speech-pause detection