Search

Your search keyword '"Chair of Complex and Intelligent Systems (CIS)"' showing total 15 results
Start Over Author "Chair of Complex and Intelligent Systems (CIS)" Topic 02 engineering and technology
15 results on '"Chair of Complex and Intelligent Systems (CIS)"'

1. Spectral and cepstral audio noise reduction techniques in speech emotion recognition

Author

Björn Schuller, Jouni Pohjalainen, Zixing Zhang, Fabien Ringeval, Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Department of Computing [London], Biomedical Image Analysis Group [London] (BioMedIA), Imperial College London-Imperial College London, This work has been supported by the European Community’s Seventh Framework Programme through the ERC Starting Grant No. 338164 (iHEARu)., and European Project: 338164,EC:FP7:ERC,ERC-2013-StG,IHEARU(2014)

Subjects
Speech emotion recognition, Computer science, Noise reduction, Speech recognition, 02 engineering and technology, Signal, 030507 speech-language pathology & audiology, 03 medical and health sciences, Cepstrum, 0202 electrical engineering, electronic engineering, information engineering, Emotion recognition, Valence (psychology), Denoising, Audio signal, business.industry, 020206 networking & telecommunications, Pattern recognition, Support vector machine, Variable (computer science), Computer Science::Sound, [INFO.INFO-IR]Computer Science [cs]/Information Retrieval [cs.IR], Artificial intelligence, ddc:004, 0305 other medical science, business, [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing
Abstract

International audience; Signal noise reduction can improve the performance of machine learning systems dealing with time signals such as audio. Real-life applicability of these recognition technologies requires the system to uphold its performance level in variable, challenging conditions such as noisy environments. In this contribution, we investigate audio signal denoising methods in cepstral and log-spectral domains and compare them with common implementations of standard techniques. The different approaches are first compared generally using averaged acoustic distance metrics. They are then applied to automatic recognition of spontaneous and natural emotions under simulated smartphone-recorded noisy conditions. Emotion recognition is implemented as support vector regression for continuous-valued prediction of arousal and valence on a realistic multimodal database. In the experiments, the proposed methods are found to generally outperform standard noise reduction algorithms.

Published
2020
Full Text
View/download PDF

2. An array of physical sensors and an adaptive regression strategy for emotion recognition in a noisy scenario

Author

Corrado Di Natale, Francesco Mosciano, Björn Schuller, Fabien Ringeval, Arianna Mencattini, Eugenio Martinelli, Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Chair of Complex and Intelligent Systems (CIS), and Universität Passau [Passau]

Subjects
Computer science, Speech recognition, 02 engineering and technology, Coatings and Films, Adaptive regression strategy, Electronic, Settore ING-INF/07 - Misure Elettriche e Elettroniche, 0202 electrical engineering, electronic engineering, information engineering, sort, [INFO]Computer Science [cs], Optical and Magnetic Materials, Electrical and Electronic Engineering, Adaptation (computer science), Instrumentation, Sensor array, ComputingMilieux_MISCELLANEOUS, Facial expression, Metals and Alloys, Estimator, 021001 nanoscience & nanotechnology, Condensed Matter Physics, Surfaces, Coatings and Films, Electronic, Optical and Magnetic Materials, Surfaces, Identification (information), Proof of concept, 020201 artificial intelligence & image processing, Emotion recognition, Affine transformation, Noise (video), 0210 nano-technology, 2506
Abstract

Several studies demonstrate that since emotions are spontaneously manifested through different measurable quantities (e.g. vocal and facial expressions), this makes possible a sort of automatic estimation of emotion from objective measurements. However, the reliability of such estimations is strongly influenced by the availability of the different sensor modalities used to monitor the affective status of a subject, and furthermore the extraction of objective parameters is sometime thwarted in a noisy and disturbed environment. This paper introduces a personalized emotion estimation based on a heterogeneous array of physical sensors for the measurement of vocal, facial, and physiological (electro-cardiogram and electro-dermal) activities. As a proof of concept, changes in the levels of both emotion reactiveness and pleasantness are estimated under critical operative conditions. The estimator model takes advantage from the time-varying selection of the most relevant non-spurious sensors features and the adaptation of the k-nearest neighbour paradigm to the continuous identification of the most affine model templates. The model, once trained, demonstrated to autonomously embed new sensorial input and adapt to unwanted/unpredicted sensor noise or emotion alteration. The proposed approach has been successfully tested on the RECOLA database, a multi-sensorial corpus of spontaneous emotional interactions in French.

Published
2017
Full Text
View/download PDF

3. Affective and Behavioural Computing: Lessons Learnt from the First Computational Paralinguistics Challenge

Author

Anton Batliner, Marcello Mortillaro, Stefan Steidl, Fabien Ringeval, Felix Weninger, Florian Eyben, Mohamed Chetouani, Klaus R. Scherer, Yue Zhang, Alessandro Vinciarelli, Erik Marchi, Björn Schuller, Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Technische Universität Munchen - Université Technique de Munich [Munich, Allemagne] (TUM), Institut d'électronique fondamentale (IEF), Université Paris-Sud - Paris 11 (UP11)-Centre National de la Recherche Scientifique (CNRS), Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Associate Institute for Signal Processing [Munich], Institut Dalle Molle d'Intelligence Artificielle Perceptive' (IDIAP), Dalle Molle Foundation, Department of Psychology, University of Geneva [Switzerland], Laboratoire des Instruments et Systèmes d'Ile de France (LISIF), Université Pierre et Marie Curie - Paris 6 (UPMC), Sorbonne Université (SU), Institut des Systèmes Intelligents et de Robotique (ISIR), and Sorbonne Université (SU)-Centre National de la Recherche Scientifique (CNRS)

Subjects
Conflict, Computer science, Feature vector, media_common.quotation_subject, Autism, 02 engineering and technology, 01 natural sciences, Theoretical Computer Science, Laughter, [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], 0103 physical sciences, Realm, 0202 electrical engineering, electronic engineering, information engineering, [INFO]Computer Science [cs], Computational Paralinguistics, Challenge, Survey, 010301 acoustics, media_common, Emotion, 020206 networking & telecommunications, ddc:128.37, Human-Computer Interaction, Social Signals, ddc:004, Software, Cognitive psychology
Abstract

In this article, we review the INTERSPEECH 2013 Computational Paralinguistics ChallengE (ComParE) – the first of its kind– in light of the recent developments in affective and behavioural computing. The impact of the first ComParE instalment is manifold: first, it featured various new recognition tasks including social signals such as laughter and fillers, conflict in dyadic group discussions, and atypical communication due to pervasive developmental disorders, as well as enacted emotion; second, it marked the onset of the ComParE, subsuming all tasks investigated hitherto within the realm of computational paralinguistics; finally, besides providing a unified test-bed under well-defined and strictly comparable conditions, we present the definite feature vector used for computation of the baselines, thus laying the foundation for a successful series of follow-up Challenges. Starting with a review of the preceding INTERSPEECH Challenges, we present the four Sub-Challenges of ComParE 2013. In particular, we provide details of the Challenge databases and a meta-analysis by conducting experiments of logistic regression on single features and evaluating the performances achieved by the participants.

Published
2019

4. Summary for AVEC 2018: Bipolar Disorder and Cross-Cultural Affect Recognition

Author

Michel Valstar, Roddy Cowie, Maja Pantic, Fabien Ringeval, Björn Schuller, Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Faculty of Engineering [Nottingham], University of Nottingham, UK (UON), Department of psychology, Queen's University [Belfast] (QUB), Department of Computing [London], Biomedical Image Analysis Group [London] (BioMedIA), and Imperial College London-Imperial College London

Subjects
Cross-Cultural Emotion, Bipolar Disorder, Computer science, 020207 software engineering, 02 engineering and technology, medicine.disease, Affect (psychology), language.human_language, Conjunction (grammar), German, ComputingMethodologies_PATTERNRECOGNITION, Structured interview, Affective Computing, 0202 electrical engineering, electronic engineering, information engineering, medicine, language, Cross-cultural, 020201 artificial intelligence & image processing, [INFO]Computer Science [cs], Bipolar disorder, ddc:004, Affective computing, Cognitive psychology
Abstract

International audience; The eighth AudioVisual Emotion Challenge and workshop AVEC 2018 was held in conjunction with ACM Multimedia'18. This year, the AVEC series addressed major novelties with three distinct sub-challenges: bipolar disorder classification, cross-cultural dimensional emotion recognition, and emotional label generation from individual ratings. The Bipolar Disorder Sub-challenge was based on a novel dataset of structured interviews of patients suffering from bipolar disorder (BD corpus), the Cross-cultural Emotion Sub-challenge relied on an extension of the SEWA dataset, which includes human-human interactions recorded 'in-the-wild' for the German and the Hungarian cultures, and the Gold-standard Emotion Sub-challenge was based on the RECOLA dataset, which was previously used in the AVEC series for emotion recognition. In this summary, we mainly describe participation and conditions of the AVEC Challenge.

Published
2018
Full Text
View/download PDF

5. AVEC 2018 Workshop and Challenge: Bipolar Disorder and Cross-Cultural Affect Recognition

Author

Roddy Cowie, Albert Ali Salah, Adrien Michaud, Hüseyin Güleç, Denis Lalanne, Björn Schuller, Maximilian Schmitt, Shahin Amiriparian, Michel Valstar, Nicholas Cummins, Elvan Ciftci, Maja Pantic, Heysem Kaya, Fabien Ringeval, Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Department of psychology, Queen's University [Belfast] (QUB), Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Department of Computer Engineering [Istanbul], Boğaziçi University [Istanbul], Faculty of Engineering [Nottingham], University of Nottingham, UK (UON), Human-IST Research Center, Universite de Fribourg, Universiteit van Amsterdam (UvA), Department of Computing [London], Biomedical Image Analysis Group [London] (BioMedIA), Imperial College London-Imperial College London, and Boǧaziçi üniversitesi = Boğaziçi University [Istanbul]

Subjects
Computer science, Event (computing), education, Information processing, 02 engineering and technology, Affect (psychology), medicine.disease, 020204 information systems, Test set, 0202 electrical engineering, electronic engineering, information engineering, Benchmark (computing), medicine, Cross-cultural, 020201 artificial intelligence & image processing, [INFO]Computer Science [cs], Bipolar disorder, ddc:004, Affective computing, Cognitive psychology
Abstract

International audience; The Audio/Visual Emotion Challenge and Workshop (AVEC 2018) "Bipolar disorder, and cross-cultural affect recognition" is the eighth competition event aimed at the comparison of multimedia processing and machine learning methods for automatic audiovisual health and emotion analysis, with all participants competing strictly under the same conditions. The goal of the Challenge is to provide a common benchmark test set for multimodal information processing and to bring together the health and emotion recognition communities, as well as the audiovisual processing communities, to compare the relative merits of various approaches to health and emotion recognition from real-life data. This paper presents the major novelties introduced this year, the challenge guidelines, the data used, and the performance of the baseline systems on the three proposed tasks: bipolar disorder classification, cross-cultural dimensional emotion recognition, and emotional label generation from individual ratings, respectively.

Published
2018
Full Text
View/download PDF

6. Towards Conditional Adversarial Training for Predicting Emotions from Speech

Author

Jing Han, Zhao Ren, Björn Schuller, Zixing Zhang, Fabien Ringeval, Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), and Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])

Subjects
Artificial neural network, conditional adversarial training, Computer science, business.industry, Feature extraction, generative adversarial network, Image processing, 02 engineering and technology, 010501 environmental sciences, Machine learning, computer.software_genre, 01 natural sciences, Arousal, Adversarial system, 0202 electrical engineering, electronic engineering, information engineering, 020201 artificial intelligence & image processing, [INFO]Computer Science [cs], Artificial intelligence, Emotion recognition, Valence (psychology), ddc:004, business, computer, 0105 earth and related environmental sciences
Abstract

International audience; Motivated by the encouraging results recently obtained by genera-tive adversarial networks in various image processing tasks, we propose a conditional adversarial training framework to predict dimensional representations of emotion, i. e., arousal and valence, from speech signals. The framework consists of two networks, trained in an adversarial manner: The first network tries to predict emotion from acoustic features, while the second network aims at distinguishing between the predictions provided by the first network and the emotion labels from the database using the acoustic features as conditional information. We evaluate the performance of the proposed conditional adversarial training framework on the widely used emotion database RECOLA. Experimental results show that the proposed training strategy outperforms the conventional training method, and is comparable with, or even superior to other recently reported approaches, including deep and end-to-end learning.

Published
2018

7. Introduction to the Special Section on Multimedia Computing and Applications of Socio-Affective Behaviors in the wild

Author

Björn Schuller, Roddy Cowie, Michel Valstar, Maja Pantic, Jonathan Gratch, Fabien Ringeval, Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Faculty of Engineering [Nottingham], University of Nottingham, UK (UON), Institute for Creative Technologies (ICT), University of Southern California (USC), Department of psychology, Queen's University [Belfast] (QUB), Department of Computing [London], Biomedical Image Analysis Group [London] (BioMedIA), and Imperial College London-Imperial College London

Subjects
Multimedia, Computer Networks and Communications, Hardware and Architecture, 0202 electrical engineering, electronic engineering, information engineering, Special section, 020201 artificial intelligence & image processing, [INFO]Computer Science [cs], 02 engineering and technology, ddc:004, Psychology, computer.software_genre, computer
Abstract

International audience; Affective Computing and Social Signal Processing are two developing fields of research that promise to revolutionise the way humans interact with multimedia and technology in general. Affective Computing is the science of creating emotionally aware systems, including the automatic analysis of affect and expressive behaviours. Whereas social signal processing addresses all verbal and non-verbal communicative signalling during social interactions, be they of an affective nature or not. Despite the scientific and technological revolutions of the last decades in these fields, there is still an urgent need to advance emotion and social behaviours recognition for multimodal interface interaction to a level where systems are able to deal with large volumes of non-prototypical naturalistic behaviour collected in ecologically valid environments, as this is exactly the type of data that applications would have to face in the real world. The first article of this special section addresses the analysis of malicious forms of communication in online social networks. Securing multimedia sharing and phishing attempts in social networks has become a critical issue with the rapidly increasing JR exploited two Twitter datasets to classify users as being human, legitimate robot or malicious robot. In the proposed method, occurrence of the terms present in the tweets are weighted by lexicon based coefficients and features are decomposed by a Discrete Wavelet Transform to feed a model based on a Random Forest classifier. An accuracy above 93% is reported for the detection of the three types of users (human, bot, malicious bot), with better results when the Tweets are related to a specific hashtag (Super Bowl event), in comparison with miscellaneous themes. The second article focuses on pose-invariant representation learning of facial expressions for emotion recognition. In the article "Spatially Coherent Feature Learning for Pose-invariant Facial Expression Recognition", Feifei Zhang, Qirong Mao, Xiangjun Shen, Yongzhao Zhan and Ming Dong, exploit a 3D pose normalisation technique that helps preserving identity information while avoiding the use of multiple models for pose-specific training and parameter tuning. In the proposed method, a sequence of key regions in a synthetic frontal face image is selected for unsupervised representation learning. Dependencies between regions of the face are specifically encoded by introducing a linkage structure over the learning-based features and the corresponding geometry information of each key region. Experimental evaluations were conducted on the BU-3DFE and SFEW datasets, and showed that the proposed model achieved superior results than others existing methods on those two datasets. We hope that this special section will bring inspirational ideas to researchers working in the domains of affective computing and/or social signal processing. We would like to thank all the authors who submitted their work to the special section, as well as the reviewers who contributed with their comments and observations to the high quality of the selected papers. We would also especially thank the editorial members of ACM TOMM, Prof. Alberto Del Bimbo and Prof. Stefano Berretti, for their support and kind help in the realisation of this special section.

Published
2018

8. AVEC 2017--Real-life depression, and affect recognition workshop and challenge

Author

Stefan Scherer, Michel Valstar, Fabien Ringeval, Roddy Cowie, Jonathan Gratch, Maximilian Schmitt, Maja Pantic, Björn Schuller, Sharon Mozgai, Nicholas Cummins, Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Imperial College London, University of Nottingham, UK (UON), Institute for Creative Technologies (ICT), University of Southern California (USC), Department of Psychology, Queen's University [Belfast] (QUB), Trinity College Dublin, Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Department of Computing [London], Biomedical Image Analysis Group [London] (BioMedIA), and Imperial College London-Imperial College London

Subjects
Event (computing), Information processing, 020206 networking & telecommunications, 02 engineering and technology, Affect (psychology), Social Signal Processing, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], ComputingMethodologies_PATTERNRECOGNITION, 13. Climate action, Test set, Affective Computing, 0202 electrical engineering, electronic engineering, information engineering, Benchmark (computing), 020201 artificial intelligence & image processing, Emotion recognition, ddc:004, Psychology, Affective computing, Social psychology, Automatic Emotion/Depression Recognition, Depression (differential diagnoses), Cognitive psychology
Abstract

International audience; The Audio/Visual Emotion Challenge and Workshop (AVEC 2017) "Real-life depression, and affect" will be the seventh competition event aimed at comparison of multimedia processing and machine learning methods for automatic audiovisual depression and emotion analysis, with all participants competing under strictly the same conditions. The goal of the Challenge is to provide a common benchmark test set for multimodal information processing and to bring together the depression and emotion recognition communities, as well as the audiovisual processing communities, to compare the relative merits of the various approaches to depression and emotion recognition from real-life data. This paper presents the novelties introduced this year, the challenge guidelines, the data used, and the performance of the baseline system on the two proposed tasks: dimensional emotion recognition (time and value-continuous), and dimensional depression estimation (value-continuous).

Published
2017
Full Text
View/download PDF

9. Speech-based Diagnosis of Autism Spectrum Condition by Generative Adversarial Network Representations

Author

Fabien Ringeval, Jun Deng, Björn Schuller, Nicholas Cummins, Maximilian Schmitt, Kun Qian, College of Engineering [Beijing], China Agricultural University (CAU), Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Universität Augsburg [Augsburg], Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), and Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])

Subjects
Computer science, Speech recognition, 02 engineering and technology, Machine learning, computer.software_genre, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], representation learning, Discriminative model, 0202 electrical engineering, electronic engineering, information engineering, medicine, Average recall, [INFO]Computer Science [cs], Training set, business.industry, 020206 networking & telecommunications, medicine.disease, automatic diagnosis, Support vector machine, Autism Spectrum Condition, Autism, 020201 artificial intelligence & image processing, Artificial intelligence, ddc:004, generative adversarial networks, business, computer, Classifier (UML), Feature learning, Generative adversarial network
Abstract

International audience; Machine learning paradigms based on child vocalisations show great promise as an objective marker of developmental disorders such as Autism. In conventional detection systems, hand-craaed acoustic features are usually fed into a discriminative classiier (e. g., Support Vector Machines); however it is well known that the accuracy and robustness of such a system is limited by the size of the associated training data. is paper explores, for the rst time, the use of feature representations learnt using a deep Genera-tive Adversarial Network (GAN) for classifying children's speech aaected by developmental disorders. A comparative evaluation of our proposed system with diierent acoustic feature sets is performed on the Child Pathological and Emotional Speech database. Key experimental results presented demonstrate that GAN based methods exhibit competitive performance with the conventional paradigms in terms of the unweighted average recall metric.

Published
2017
Full Text
View/download PDF

10. Reconstruction-error-based learning for continuous emotion recognition in speech

Author

Björn Schuller, Fabien Ringeval, Jing Han, Zixing Zhang, Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Department of Computing [London], Biomedical Image Analysis Group [London] (BioMedIA), Imperial College London-Imperial College London, This work was partially supported by the EC’s 7th Framework Programme through the ERC Starting Grant No. 338164 (iHEARu), the EU’s Horizon 2020 Programme through the Innovative Action No. 645094 (SEWA) and the Research Innovative Action No. 645378 (ARIA-VALUSPA), and by the German Federal Ministry of Education, Science, Research and Technology (BMBF) under grant agreement No. 16SV7213 (EmotAsS)., European Project: 338164,EC:FP7:ERC,ERC-2013-StG,IHEARU(2014), European Project: 645094,H2020,H2020-ICT-2014-1,SEWA(2015), and European Project: 645378,H2020,H2020-ICT-2014-1,ARIA-VALUSPA(2015)

Subjects
Computer science, business.industry, Speech recognition, Continuous emotion recognition, Feature extraction, Reconstruction error, 020207 software engineering, Pattern recognition, 02 engineering and technology, Autoencoder, Recurrent neural network, Concordance correlation coefficient, [STAT.ML]Statistics [stat]/Machine Learning [stat.ML], [INFO.INFO-IR]Computer Science [cs]/Information Retrieval [cs.IR], 0202 electrical engineering, electronic engineering, information engineering, 020201 artificial intelligence & image processing, Artificial intelligence, Emotion recognition, Valence (psychology), ddc:004, Hidden Markov model, business, Bidirectional long short-term memory
Abstract

International audience; To advance the performance of continuous emotion recognition from speech, we introduce a reconstruction-error-based (RE-based) learning framework with memory-enhanced Recurrent Neural Networks (RNN). In the framework, two successive RNN models are adopted, where the first model is used as an autoencoder for reconstructing the original features, and the second is employed to perform emotion prediction. The RE of the original features is used as a complementary descriptor, which is merged with the original features and fed to the second model. The assumption of this framework is that the system has the ability to learn its 'drawback' which is expressed by the RE. Experimental results on the RECOLA database show that the proposed framework significantly outperforms the baseline systems without any RE information in terms of Concordance Correlation Coefficient (.729 vs .710 for arousal, .360 vs .237 for valence), and also significantly overcomes other state-of-the-art methods.

Published
2017

11. Strength modelling for real-worldautomatic continuous affect recognition from audiovisual signals

Author

Björn Schuller, Fabien Ringeval, Zixing Zhang, Jing Han, Nicholas Cummins, Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Department of Computing [London], Biomedical Image Analysis Group [London] (BioMedIA), Imperial College London-Imperial College London, This work was supported by the EU’s Horizon 2020 Programme through the Innovative Action No. 645094 (SEWA) and the EC’s 7th Framework Programme through the ERC Starting Grant No. 338164 (iHEARu). We further thank the NVIDIA Corporation for their support of this research by Tesla K40-type GPU donation., European Project: 645094,H2020,H2020-ICT-2014-1,SEWA(2015), and European Project: 338164,EC:FP7:ERC,ERC-2013-StG,IHEARU(2014)

Subjects
Artificial neural network, Computer science, business.industry, Feature vector, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], 020206 networking & telecommunications, 02 engineering and technology, Machine learning, computer.software_genre, Regression, Support vector machine, Robustness (computer science), Signal Processing, 0202 electrical engineering, electronic engineering, information engineering, Leverage (statistics), 020201 artificial intelligence & image processing, Computer Vision and Pattern Recognition, Artificial intelligence, Performance improvement, business, Regression problems, computer
Abstract

International audience; Automatic continuous affect recognition from audiovisual cues is arguably one of the most active research areas in machine learning. In addressing this regression problem, the advantages of the models, such as the global-optimisation capability of Support Vector Machine for Regression and the context-sensitive capability of memory-enhanced neural networks, have been frequently explored, but in an isolated way. Motivated to leverage the individual advantages of these techniques, this paper proposes and explores a novel framework, Strength Modelling, where two models are concatenated in a hierarchical framework. In doing this, the strength information of the first model, as represented by its predictions, is joined with the original features, and this expanded feature space is then utilised as the input by the successive model. A major advantage of Strength Modelling, besides its ability to hierarchically explore the strength of different machine learning algorithms, is that it can work together with the conventional feature-and decision-level fusion strategies for multimodal affect recognition. To highlight the effectiveness and robustness of the proposed approach, extensive experiments have been carried out on two time-and value-continuous spontaneous emotion databases (RECOLA and SEMAINE) using audio and video signals. The experimental results indicate that employing Strength Modelling can deliver a significant performance improvement for both arousal and valence in the unimodal and bimodal settings. The results further show that the proposed systems is competitive or outperform the other state-of-the-art approaches, but being with a simple implementation.

Published
2017
Full Text
View/download PDF

12. Summary for AVEC 2016 – Depression, Mood, and Emotion Recognition Workshop and Challenge

Author

Michel Valstar, Maja Pantic, Jonathan Gratch, Fabien Ringeval, Roddy Cowie, Björn Schuller, Faculty of Engineering [Nottingham], University of Nottingham, UK (UON), Institute for Creative Technologies (ICT), University of Southern California (USC), Department of Computing [London], Biomedical Image Analysis Group [London] (BioMedIA), Imperial College London-Imperial College London, Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Department of psychology, Queen's University [Belfast] (QUB), University of Twente [Netherlands], The challenge in general has been generously supported by the Association for the Advancement of Affective Computing (AAAC). The research leading to these results has received funding from the EC’s 7th Framework Programme through the ERC Starting Grant No. 338164 (iHEARu), and the European Union’s Horizon 2020 Programme through the Innovative Action No. 645094 (SEWA), and the Research Innovative Action No. 645378 (ARIA-VALUSPA), and No. 688835 (DE-ENIGMA)., European Project: 338164,EC:FP7:ERC,ERC-2013-StG,IHEARU(2014), European Project: 645094,H2020,H2020-ICT-2014-1,SEWA(2015), European Project: 645378,H2020,H2020-ICT-2014-1,ARIA-VALUSPA(2015), European Project: 688835,H2020,H2020-ICT-2015,DE-ENIGMA(2016), and University of Twente

Subjects
Computer science, Depression, 02 engineering and technology, Social Signal Processing, Conjunction (grammar), 03 medical and health sciences, 0302 clinical medicine, Mood, [STAT.ML]Statistics [stat]/Machine Learning [stat.ML], 13. Climate action, [INFO.INFO-IR]Computer Science [cs]/Information Retrieval [cs.IR], Affective Computing, 0202 electrical engineering, electronic engineering, information engineering, 020201 artificial intelligence & image processing, Emotion recognition, ddc:004, Affective computing, 030217 neurology & neurosurgery, Depression (differential diagnoses), Cognitive psychology
Abstract

International audience; The sixth AudioVisual Emotion Challenge and workshop AVEC 2016 was held in conjunction ACM Multimedia'16. This year the AVEC series addresses two distinct sub-challenges, multi-modal emotion recognition and audiovisual depression detection. Both sub-challenges are in a way a return to AVEC's past editions: the emotion sub-challenge is based on the same dataset as the one used in AVEC 2015, and depression analysis was previously addressed in AVEC 2013/2014. In this summary, we mainly describe participation and its conditions.

Published
2016
Full Text
View/download PDF

13. Facing Realism in Spontaneous Emotion Recognition from Speech: Feature Enhancement by Autoencoder with LSTM Neural Networks

Author

Fabien Ringeval, Björn Schuller, Zixing Zhang, Erik Marchi, Jun Deng, Jing Han, Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Department of Computing [London], Biomedical Image Analysis Group [London] (BioMedIA), Imperial College London-Imperial College London, The research leading to these results was supported by the EC’s 7th Framework Programme through the ERC Starting Grant No. 338164 (iHEARu), the EU’s Horizon 2020 Programme through the Innovative Action No. 644632 (MixedEmotions), No. 645094 (SEWA), and the Research Innovative Action No. 645378 (ARIA-VALUSPA), and by the German Federal Ministry of Education, Science, Research and Technology (BMBF) under grant agreement #16SV7213 (EmotAsS). We further thank the NVIDIA Corporation for their support of this research by Tesla K40-type GPU donation., European Project: 338164,EC:FP7:ERC,ERC-2013-StG,IHEARU(2014), European Project: 644632,H2020,H2020-ICT-2014-1,MixedEmotions(2015), European Project: 645094,H2020,H2020-ICT-2014-1,SEWA(2015), and European Project: 645378,H2020,H2020-ICT-2014-1,ARIA-VALUSPA(2015)

Subjects
autoencoder, Artificial neural network, Computer science, Speech recognition, 020207 software engineering, 02 engineering and technology, spontaneous speech, LSTM Neural Networks, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], Autoencoder, Arousal, additive and convolutional noises, Support vector machine, Face (geometry), [INFO.INFO-IR]Computer Science [cs]/Information Retrieval [cs.IR], emotion recognition, [INFO.INFO-SD]Computer Science [cs]/Sound [cs.SD], 0202 electrical engineering, electronic engineering, information engineering, Feature (machine learning), 020201 artificial intelligence & image processing, feature enhancement, Noise (video), Valence (psychology), ddc:004
Abstract

International audience; During the last decade, speech emotion recognition technology has matured well enough to be used in some real-life scenarios. However, these scenarios require an almost silent environment to not compromise the performance of the system. Emotion recognition technology from speech thus needs to evolve and face more challenging conditions, such as environmental additive and convolutional noises, in order to broaden its applicability to real-life conditions. This contribution evaluates the impact of a front-end feature enhancement method based on an autoencoder with long short-term memory neural networks, for robust emotion recognition from speech. Support Vector Regression is then used as a back-end for time- and value-continuous emotion prediction from enhanced features. We perform extensive evaluations on both non-stationary additive noise and convolutional noise, on a database of spontaneous and natural emotions. Results show that the proposed method significantly outperforms a system trained on raw features, for both arousal and valence dimensions, while having almost no degradation when applied to clean speech.

Published
2016
Full Text
View/download PDF

14. Continuous Estimation of Emotions in Speech by Dynamic Cooperative Speaker Models

Author

Fabien Ringeval, Corrado Di Natale, Arianna Mencattini, Eugenio Martinelli, Björn Schuller, Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP ), Laboratoire d'Informatique de Grenoble (LIG ), Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019])-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes [2016-2019] (UGA [2016-2019]), Chair of Complex and Intelligent Systems (CIS), and Universität Passau [Passau]

Subjects
Speech emotion recognition, Speech recognition, 02 engineering and technology, Machine learning, computer.software_genre, cooperative regression model, Data modeling, Arousal, 030507 speech-language pathology & audiology, 03 medical and health sciences, 0202 electrical engineering, electronic engineering, information engineering, Recognition system, Settore ING-INF/07 - Misure Elettriche e Elettroniche, [INFO]Computer Science [cs], Emotion recognition, Valence (psychology), business.industry, Modular architecture, Speaker recognition, Human-Computer Interaction, naturalistic emotional display, 020201 artificial intelligence & image processing, Artificial intelligence, ddc:004, 0305 other medical science, business, Psychology, computer, Software
Abstract

International audience; Research on automatic emotion recognition from speech has recently focused on the prediction of time-continuous dimensions (e. g., arousal and valence) of spontaneous and realistic expressions of emotion, as found in real-life interactions. However, the automatic prediction of such emotions poses several challenges, such as the subjectivity found in the definition of a gold-standard from a pool of raters and the issue of data scarcity in training models. In this work, we introduce a novel emotion recognition system, based on ensembles of single-speaker-regression-models. The estimation of emotion is provided by combining a subset of the initial pool of single-speaker-regression-models selecting those that are most concordant among them. The proposed approach allows the addition or removal of speakers from the ensemble without the necessity to rebuild the entire recognition system. The simplicity of this aggregation strategy, coupled with the flexibility assured by the modular architecture, and the promising results observed on the RECOLA database highlight the potential implications of the proposed method in a real-life scenario and in particular in web-based applications.

Published
2016
Full Text
View/download PDF

15. The INTERSPEECH 2013 computational paralinguistics challenge: Social signals, conflict, emotion, autism

Author

Hugues Salamin, Erik Marchi, Mohamed Chetouani, Stefan Steidl, Samuel Kim, Fabio Valente, Felix Weninger, Klaus R. Scherer, Fabien Ringeval, Alessandro Vinciarelli, Florian Eyben, Anna Polychroniou, Anton Batliner, Marcello Mortillaro, Björn Schuller, Chetouani, Mohamed, Universität Augsburg [Augsburg], Institut Dalle Molle d'Intelligence Artificielle Perceptive' (IDIAP), Dalle Molle Foundation, Department of Psychology, University of Geneva [Switzerland], Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole (GETALP), Laboratoire d'Informatique de Grenoble (LIG), Université Pierre Mendès France - Grenoble 2 (UPMF)-Université Joseph Fourier - Grenoble 1 (UJF)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Institut National Polytechnique de Grenoble (INPG)-Centre National de la Recherche Scientifique (CNRS)-Université Pierre Mendès France - Grenoble 2 (UPMF)-Université Joseph Fourier - Grenoble 1 (UJF)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Institut National Polytechnique de Grenoble (INPG)-Centre National de la Recherche Scientifique (CNRS), Institut des Systèmes Intelligents et de Robotique (ISIR), Université Pierre et Marie Curie - Paris 6 (UPMC)-Centre National de la Recherche Scientifique (CNRS), Sorbonne Université (SU), Technische Universität Munchen - Université Technique de Munich [Munich, Allemagne] (TUM), Associate Institute for Signal Processing [Munich], Chair of Complex and Intelligent Systems (CIS), Universität Passau [Passau], Université Pierre Mendès France - Grenoble 2 (UPMF)-Université Joseph Fourier - Grenoble 1 (UJF)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP)-Institut National Polytechnique de Grenoble (INPG)-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA)-Université Pierre Mendès France - Grenoble 2 (UPMF)-Université Joseph Fourier - Grenoble 1 (UJF)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP)-Institut National Polytechnique de Grenoble (INPG)-Centre National de la Recherche Scientifique (CNRS)-Université Grenoble Alpes (UGA), Technische Universität München [München] (TUM), and Université de Genève = University of Geneva (UNIGE)

Subjects
[INFO.INFO-AI] Computer Science [cs]/Artificial Intelligence [cs.AI], Conflict, Computer science, [INFO.INFO-TS] Computer Science [cs]/Signal and Image Processing, media_common.quotation_subject, Speech recognition, Autism, [SCCO.COMP]Cognitive science/Computer science, 02 engineering and technology, Task (project management), [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], Laughter, [INFO.INFO-CV] Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, [SCCO.COMP] Cognitive science/Computer science, 0502 economics and business, 0202 electrical engineering, electronic engineering, information engineering, medicine, [INFO.INFO-RB]Computer Science [cs]/Robotics [cs.RO], Computational Paralinguistics, Index Terms: Computational Paralinguistics, Challenge, media_common, Emotion, [SCCO.NEUR]Cognitive science/Neuroscience, 05 social sciences, [INFO.INFO-RB] Computer Science [cs]/Robotics [cs.RO], [SCCO.NEUR] Cognitive science/Neuroscience, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], medicine.disease, ddc:128.37, Social Signals, [SCCO.PSYC]Cognitive science/Psychology, [SCCO.PSYC] Cognitive science/Psychology, 050211 marketing, 020201 artificial intelligence & image processing, ddc:004, Cognitive psychology
Abstract

The INTERSPEECH 2013 Computational Paralinguistics Challenge provides for the first time a unified test-bed for Social Signals such as laughter in speech. It further introduces conflict in group discussions as a new task and deals with autism and its manifestations in speech. Finally, emotion is revisited as task, albeit with a broader range of overall twelve enacted emotional states. In this paper, we describe these four Sub-Challenges, their conditions, baselines, and a new feature set by the openSMILE toolkit, provided to the participants. Index Terms: Computational Paralinguistics, Challenge, Social Signals, Conflict, Emotion, Autism

Catalog

Books, media, physical & digital resources
See catalog results