Faculty of Information Technology, BUT

Publication Details

Application of speaker- and language identification state-of-the-art techniques for emotion recognition

KOCKMANN Marcel, BURGET Lukáš and ČERNOCKÝ Jan. Application of speaker- and language identification state-of-the-art techniques for emotion recognition. Speech Communication, vol. 53, no. 9, pp. 1172-1185. ISSN 0167-6393. Available from: http://pdn.sciencedirect.com/science?_ob=MiamiImageURL&_cid=271578&_user=640830&_pii=S0167639311000082&_check=y&_origin=search&_zone=rslt_list_item&_coverDate=2011-12-31&wchp=dGLbVlS-zSkWz&md5=2a79c3d171cd13a3689408115666e2ef/1-s2.0-S0167639311000082-main
Czech title
Použití aktuálních technik pro identifikaci řečníka a jazyka v rozpoznávání emocí
Type
journal article
Language
english
Authors
URL
Keywords
Emotion recognition; Gaussian mixture models; Maximum-mutual-information; Intersession variability compensation; Score-level fusion
Abstract
Authors of this article show that feature extraction and statistical modeling methods that are usually used in speaker and language recognition can be successfully used for emotion recognition as well.
Annotation
This article describes our efforts of transferring feature extraction and statistical modeling techniques from the fields of speaker and language identification to the related field of emotion recognition. We give detailed insight to our acoustic and prosodic feature extraction and show how to apply Gaussian Mixture Modeling techniques on top of it. We focus on different flavors of Gaussian Mixture Models (GMMs), including more sophisticated approaches like discriminative training using Maximum-Mutual-Information (MMI) criterion and InterSession Variability (ISV) compensation. Both techniques show superior performance in language and speaker identification. Furthermore, we combine multiple system outputs by score-level fusion to exploit the complementary information in diverse systems. Our proposal is evaluated with several experiments on the FAU Aibo Emotion Corpus containing non-acted spontaneous emotional speech. Within the Interspeech 2009 Emotion Challenge we could achieve the best results for the 5-class task of the Open Performance Sub-Challenge with an unweighted average recall of 41.7%. Further additional experiments on the acted Berlin Database of Emotional Speech show the capability of intersession variability compensation for emotion recognition.
Published
2011
Pages
1172-1185
Journal
Speech Communication, vol. 53, no. 9, ISSN 0167-6393
Book
Speech Communication
Publisher
Elsevier Science
DOI
BibTeX
@ARTICLE{FITPUB9676,
   author = "Marcel Kockmann and Luk\'{a}\v{s} Burget and Jan \v{C}ernock\'{y}",
   title = "Application of speaker- and language identification state-of-the-art techniques for emotion recognition",
   pages = "1172--1185",
   booktitle = "Speech Communication",
   journal = "Speech Communication",
   volume = 53,
   number = 9,
   year = 2011,
   publisher = "Elsevier Science",
   ISSN = "0167-6393",
   doi = "10.1016/j.specom.2011.01.007",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/9676"
}
Back to top