Building robust emotion recognition system on heterogeneous speech databases

Yoon, Won-Jung; Park, Kyu-Sik

doi:10.1109/tce.2011.5955217

Cited by 21 publications

(7 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Also, feature generation (Schuller et al, 2006b) can be employed. Recently, more optimisation for data representation is utilised such as instance balancing (e. g., by SMOTE (Schuller et al, 2009b) or similar techniques) or instance selection (Erdem et al, 2010;Schuller et al, 2011f), and more systematic classifier and regressor optimisation and tailored architectures (hierarchical (Yoon and Park, 2011) and hybrid (Schuller et al, 2004) or ensembles (Schuller et al, 2005;Schwenker et al, 2010)). The Challenge reported above has also shown an increasing trend towards fusion of multiple systems.…”

Section: Ten Recent and Future Trendsmentioning

confidence: 99%

Paralinguistics in speech and language—State-of-the-art and the challenge

Schuller

Steidl

Batliner

et al. 2013

Computer Speech & Language

263

159

View full text Add to dashboard Cite

Paralinguistic analysis is increasingly turning into a mainstream topic in speech and language processing. This article aims to provide a broad overview of the constantly growing field by defining the field, introducing typical applications, presenting exemplary resources, and sharing a unified view of the chain of processing. It then presents the first broader Paralinguistic Challenge organised at INTERSPEECH 2010 by the authors including a historical overview of the Challenge tasks of recognising age, gender, and affect, a summary of methods used by the participants, and their results. In addition, we present the new benchmark obtained by fusion of participants' predictions and conclude by discussing ten recent and emerging trends in the analysis of paralinguistics in speech and language.

show abstract

Section: Ten Recent and Future Trendsmentioning

confidence: 99%

Paralinguistics in speech and language—State-of-the-art and the challenge

Schuller

Steidl

Batliner

et al. 2013

Computer Speech & Language

263

159

View full text Add to dashboard Cite

show abstract

“…However, pitch is also the most genderdependent feature [12]. Yoon [13] had taken gender differences into consideration while applying speech emotion recognition.…”

Section: Speech Emotion Recognitionmentioning

confidence: 99%

Techniques in Pattern Recognition for School Bullying Prevention: Review and Outlook

Liang¹,

Ferdinando²,

Alasaarela³

2014

JPRR

View full text Add to dashboard Cite

School bullying is a serious problem among teenagers. With the development of sensor technology and pattern recognition algorithms, several approaches for detecting school bullying have been developed, namely speech emotion recognition, mental stress recognition, and activity recognition. This paper reviews some related work and makes some comparisons among these three aspects. The paper analyzes commonly used features and classifiers, and describes some examples. The Gaussian Mixture Model and the Double Threshold classifiers provided high accuracies in many experiments. By using a combined architecture of classifiers, the results could be further improved. According to the results of the experiments, the six basic emotions, high mental stress and irregular movements can be recognized with high accuracies. So the three types of pattern recognition can be used for school bullying detection effectively. And these techniques can be used on consumer devices such as smartphones to protect teenagers.

show abstract

“…Besides, there is an increasing trend towards fusion of multiple systems, as has been evident in the sequence of paralinguistic challenges [76][77][78]. Fusion can be applied to classifier decisions in hierarchical [44,101], hybrid [75] or ensemble architectures [73,86]; at an even higher level, fusing the output of entire recognition systems can successfully exploit their complementarity; for instance, majority voting among the systems from the best participants in the Interspeech 2009 Emotion Challenge yields the best result reported so far on the challenge corpus [71].…”

Section: More Optimisationmentioning

confidence: 99%

Ten Recent Trends in Computational Paralinguistics

Schuller

Weninger

2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

The field of computational paralinguistics is currently emerging from loosely connected research in speech analysis, including speaker classification and emotion recognition. Starting from a broad perspective on the state-of-the-art in this field, we combine these facts with a bit of 'tea leaf reading' to identify ten trends that might characterise the next decade of research: taking into account more tasks and task interdependencies, modelling paralinguistic information in the continuous domain, agglomerating and evaluating on large amounts of heterogeneous data, exploiting more and more types of features, fusing linguistic and non-linguistic phenomena, devoting more effort to optimisation of the machine learning aspects, standardising the whole processing chain, addressing robustness and security of systems, proceeding to evaluation in real-life conditions, and finally overcoming cross-language and crosscultural barriers. We conjecture that following these trends we will see an increase in the 'social competence' of tomorrow's technical systems.

show abstract

Building robust emotion recognition system on heterogeneous speech databases

Cited by 21 publications

References 6 publications

Paralinguistics in speech and language—State-of-the-art and the challenge

Paralinguistics in speech and language—State-of-the-art and the challenge

Techniques in Pattern Recognition for School Bullying Prevention: Review and Outlook

Ten Recent Trends in Computational Paralinguistics

Contact Info

Product

Resources

About