Recognition of Emotions in German Speech Using Gaussian Mixture Models
- 856 Downloads
The contribution describes experiments with recognition of emotions in German speech signal based on the same principle as recognition of speakers. The most robust algorithm for speaker recognition is based on Gaussian Mixture Models (GMM). We examine three parameter sets: the first contains suprasegmental features, in the second are segmental features and the last is a combination of the two previous parameter sets. Further we want to explore the dependency of the classification accuracy on the number of GMM model components. The aim of this contribution is a recommendation for the number of GMM components and the optimal selection of speech parameters for emotion recognition in German speech.
Keywordsspeech emotions emotion recognition Gaussian mixture models
Unable to display preview. Download preview PDF.
- 2.Burkhardt, F., Paeschke, A., Rolfes, M., Sendlmeier, W., Weiss, B.: A Database of German Emotional Speech. In: Proc. Interspeech 2005, Lisbon, Portugal, September 4-8 (2005)Google Scholar
- 3.Truong, K.P., Leeuven, D.A.: An ‘open-set’ detection evaluation methology for automatic emotion recognition in speech. In: ParaLing 2007: Workshop on Paralinguistic Speech - between models and data, Saarbrücken, Germany (2007)Google Scholar
- 4.Morrison, D., Wang, R., De Silva, L.C.: Ensemble methods for spoken emotion recognition in call-centers. Speech Communication 49 (2007)Google Scholar
- 5.Sjölander, K., Beskow, J.: Wavesurfer, http://www.speech.kth.se/wavesurfer/
- 6.Brookes, M.: VOICEBOX: Speech Processing Toolbox for MATLAB, http://www.ee.ic.ac.uk/hp/staff/dmb/voicebox/voicebox.html