Speech Prosody 2010

Chicago, IL, USA
May 10-14, 2010

Automatic Classification of Emotions via Global and Local Prosodic Features on a Multilingual Emotional Database

Antonio Origlia (1), Vincenzo Galatà (2), Bogdan Ludusan (1)

(1) Natural Language Processing Group, Department of Physical Sciences, “Federico II” UniveNaples, Italy Naples, Italy
(2) Laboratory of Phonetics, Department of Linguistics, University of Calabria, Italy

In this paper we introduce the €motion database, a multilingual emotional database consisting of emotional sentences elicited in four European languages: Italian, French, English and German. Along with this, a new set of features, containing both global and local prosodic features, for automatic classification of emotions is presented and their appropriateness for this task tested. The results obtained using these features on a monolingual emotional database are comparable with the state of the art results previously obtained.

Index Terms: automatic emotions classification, multilingual emotional database, prosodic features

Full Paper

Bibliographic reference.  Origlia, Antonio / Galatà, Vincenzo / Ludusan, Bogdan (2010): "Automatic classification of emotions via global and local prosodic features on a multilingual emotional database", In SP-2010, paper 213.