Memorias de investigación
Communications at congresses:
Acoustic Emotion Recognition using Dynamic Bayesian Networks and Multi-Space Distributions
Year:2009
Research Areas
  • Artificial intelligence,
  • Electronic circuits,
  • Electronic devices
Information
Abstract
In this paper we describe the acoustic emotion recognition system built at the Speech Technology Group of the Universidad Politecnica de Madrid (Spain) to participate in the INTERSPEECH 2009 Emotion Challenge. Our proposal is based on the use of a Dynamic Bayesian Network (DBN) to deal with the temporal modelling of the emotional speech information. The selected features (MFCC, F0, Energy and their variants) are modelled as different streams, and the F0 related ones are integrated under a Multi Space Distribution (MSD) framework, to properly model its dual nature (voiced/unvoiced). Experimental evaluation on the challenge test set, show a 67.06% and 38.24% of unweighted recall for the 2 and 5-classes tasks respectively. In the 2-class case, we achieve similar results compared with the baseline, with a considerable less number of features. In the 5-class case, we achieve a statistically significant 6.5% relative improvement. Index Terms: automatic emotion recognition, multi-space probability distribution, dynamic bayesian networks, emotion challenge
International
Si
Congress
10th Annual Conference of the International Speech Communication Association, Interspeech 2009
960
Place
Brighton UK.
Reviewers
Si
ISBN/ISSN
1990-9772
Start Date
06/09/2009
End Date
10/09/2009
From page
336
To page
339
Proceedings of 10th Annual Conference of the International Speech Communication Association, Interspeech 2009
Participants
Research Group, Departaments and Institutes related
  • Creador: Grupo de Investigación: Grupo de Tecnología del Habla
  • Departamento: Ingeniería Electrónica