Automatic multi-lingual arousal detection from voice applied to real product testing applications
Autor: | Björn Schuller, Matthias Unfried, Florian Eyben, Gerhard Hagerer |
---|---|
Jazyk: | angličtina |
Rok vydání: | 2020 |
Předmět: |
Computer science
business.industry Speech recognition 05 social sciences Feature extraction Product testing Novelty Speech corpus computer.software_genre 01 natural sciences 050105 experimental psychology Arousal Recurrent neural network 0103 physical sciences 0501 psychology and cognitive sciences Artificial intelligence ddc:004 Emotional arousal business 010301 acoustics computer Natural language processing |
Zdroj: | ICASSP |
Popis: | A method is presented which applies Long Short-Term Memory Recurrent Neural Networks on real market-research voice recordings in order to automatically predict emotional arousal from speech. While most previous work has dealt with evaluations of algorithms within the same speech corpus, the novelty of this paper lies in an extensive evaluation across corpora and languages. The approach is evaluated on seven large data sets collected in real tests of TV commercials and new product concepts across four languages. We observe excellent performance within and between the different corpora when compared against the gold standard of arousal ratings by human annotators. Even in the cross-language validation the models show good performance which almost reaches human rater agreement. |
Databáze: | OpenAIRE |
Externí odkaz: |