Cross-Corpus Acoustic Emotion Recognition: Variances and Strategies (Extended 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. 1.8k individual evaluations in total indicate the crucial performance inferiority of inter- to intra-corpus testing.
| Author: | Björn Schuller, Bogdan Vlasenko, Florian Eyben, Martin Wöllmer, André StuhlsatzORCiDGND, Andreas Wendemuth, Gerhard Rigoll |
|---|---|
| open access: | Bronze - frei zugänglich aber ohne Lizenzhinweis, d.h. keine Weitergabe |
| Fachbereich/Einrichtung: | Hochschule Düsseldorf / Fachbereich - Maschinenbau und Verfahrenstechnik |
| Document Type: | Conference Proceeding |
| Year of Completion: | 2015 |
| Language of Publication: | English |
| Parent Title (English): | In Proc. 6th biannual Conference on Affective Computing and Intelligent Interaction (ACII 2015), AAAC, |
| Related URL: | https://web.archive.org/web/20250716075551/https://publications.idiap.ch/downloads/papers/2015/Schuller_ACII2015_2015.pdf |
| Dewey Decimal Classification: | 6 Technik, Medizin, angewandte Wissenschaften / 62 Ingenieurwissenschaften / 620 Ingenieurwissenschaften und zugeordnete Tätigkeiten |
| Licence (German): | Es gilt das deutsche Urheberrecht |
| Release Date: | 2025/07/16 |
