5th International Conference on Spoken Language Processing

Sydney, Australia
November 30 - December 4, 1998

Towards an Automatic Classification of Emotions in Speech

N. Amir, S. Ron

Center for Technological Education Holon, Israel

In this paper we discuss a method for extracting emotional state from the speech signal. We describe a methodology for obtaining emotive speech. and a method for identifying the emotion present in the signal. This method is based on analysis of the signal over sliding windows. and extracting a representative parameter set. A set of basic emotions is defined, and for each such emotion a reference point is computed. At each instant the distance of the measured parameter set from the reference points is calculated. and used to compute a fuzzy membership index for each emotion, which we term the "emotional index". Preliminary results are presented, which demonstrate the discriminative abilities of this method when applied to a number of speakers.

Full Paper

Bibliographic reference.  Amir, N. / Ron, S. (1998): "Towards an automatic classification of emotions in speech", In ICSLP-1998, paper 0199.