Cross-corpus acoustic emotion recognition: Variances and strategies (Extended abstract)

Bjorn Schuller, Bogdan Vlasenko, Florian Eyben, Martin Wollmer, Andre Stuhlsatz, Andreas Wendemuth, Gerhard Rigoll

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Scopus citations

Abstract

As the recognition of emotion from speech has matured to a degree where it becomes applicable in real-life settings, it is time for a realistic view on obtainable performances. Most studies tend to overestimation in this respect: acted data is often used rather than spontaneous data, results are reported on pre-selected prototypical data, and true speaker disjunctive partitioning is still less common than simple cross-validation. A considerably more realistic impression can be gathered by inter-set evaluation: we therefore show results employing six standard databases in a cross-corpora evaluation experiment. To better cope with the observed high variances, different types of normalization are investigated. 1.8 k individual evaluations in total indicate the crucial performance inferiority of inter- to intra-corpus testing.

Original languageEnglish
Title of host publication2015 International Conference on Affective Computing and Intelligent Interaction, ACII 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages470-476
Number of pages7
ISBN (Electronic)9781479999538
DOIs
StatePublished - 2 Dec 2015
Event2015 International Conference on Affective Computing and Intelligent Interaction, ACII 2015 - Xi'an, China
Duration: 21 Sep 201524 Sep 2015

Publication series

Name2015 International Conference on Affective Computing and Intelligent Interaction, ACII 2015

Conference

Conference2015 International Conference on Affective Computing and Intelligent Interaction, ACII 2015
Country/TerritoryChina
CityXi'an
Period21/09/1524/09/15

Keywords

  • Cross-corpus
  • Speech Emotion Recognition

Fingerprint

Dive into the research topics of 'Cross-corpus acoustic emotion recognition: Variances and strategies (Extended abstract)'. Together they form a unique fingerprint.

Cite this