Author: "Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA)" / Language: undetermined - Searchworks@Jio Institute Digital Library Search Results

Your search keyword '"Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA)"' showing total 41 results

Start Over Author "Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA)" Language undetermined

41 results on '"Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA)"'

1. Statistical learning models of early phonetic acquisition struggle with child-centered audio data

Author: Marvin Lavechin, Maureen de Seyssel, Marianne Métais, Florian Metz, Abdelrahman Mohame, Hervé Bredin, Emmanuel Dupoux, Alejandrina Cristia, Laboratoire de sciences cognitives et psycholinguistique (LSCP), Département d'Etudes Cognitives - ENS Paris (DEC), École normale supérieure - Paris (ENS-PSL), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-École normale supérieure - Paris (ENS-PSL), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-École des hautes études en sciences sociales (EHESS)-Centre National de la Recherche Scientifique (CNRS), Apprentissage machine et développement cognitif (CoML), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-École des hautes études en sciences sociales (EHESS)-Centre National de la Recherche Scientifique (CNRS)-Département d'Etudes Cognitives - ENS Paris (DEC), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-École des hautes études en sciences sociales (EHESS)-Centre National de la Recherche Scientifique (CNRS)-Inria de Paris, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria), Meta AI Research [Paris], Meta AI, Laboratoire de Linguistique Formelle (LLF - UMR7110), Centre National de la Recherche Scientifique (CNRS)-Université Paris Cité (UPCité), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT), Centre National de la Recherche Scientifique (CNRS), CIFAR (Learning in Machines and Brains), ANR-16-DATA-0004,ACLEW,Analyzing Child Language Experiences Around the World(2016), ANR-17-EURE-0017,FrontCog,Frontières en cognition(2017), ANR-19-P3IA-0001,PRAIRIE,PaRis Artificial Intelligence Research InstitutE(2019), and European Project: 101001095,ExELang
Subjects: Perceptual attunement, Domain-general, Phonetic learning, Language acquisition, Domain-specific, Child-centered long-forms, [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL]
Abstract: International audience; Infants learn their native language(s) at an amazing speed. Before they even talk, their perception adapts to the language(s) they hear. However, the mechanisms responsible for this perceptual attunement still remain unclear. A long tradition in linguistics points to the importance of specialized language mechanisms that would allow us to quickly and effortlessly learn from the language(s) we are exposed to. However, the currently dominant explanation for perceptual attunement posits that infants apply a domain-general learning mechanism consisting in learning statistical regularities from the speech stream they hear, and which may be found in learning across domains and across species. Critically, the feasibil- ity of employing purely domain-general statistical learning mechanisms has only been demonstrated with computational models on unrealistic and simplified input. This paper presents the first attempt to study perceptual attunement with 2,000 hours of ecological child-centered recordings in American English and Metropolitan French. We show that, when applied on ecologically-valid data, generic learning mecha- nisms develop a language-relevant perceptual space but fail to show evidence for perceptual attunement. It is only when supplemented with domain-specific audio filtering and augmentation mechanisms that computational models show a significant attunement to the language they have been exposed to. Hence, we conclude that, when learning from ecological audio, domain-specific mechanisms may be necessary to guide early language learning in the wild even if the learning itself is done through generic mechanisms. We anticipate our work to be a starting point for ecologically-valid computational models of perceptual attunement in other domains and species.
Published: 2023
Full Text: View/download PDF

2. PCEDNet: A Lightweight Neural Network for Fast and Interactive Edge Detection in 3D Point Clouds

Author: HimeurChems-Eddine, PellegriniThomas, BartheLoic, MelladoNicolas, LejembleThibault, PaulinMathias, Structural Models and Tools in Computer Graphics (IRIT-STORM), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), and ANR-11-BS02-0006,ALTA,Analyse des opérateurs de transport lumineux et applications(2011)
Subjects: low resource computing, FOS: Computer and information sciences, Computer science, Point cloud, 02 engineering and technology, Convolutional neural network, Edge detection, Computer Science - Graphics, 0202 electrical engineering, electronic engineering, information engineering, Segmentation, energy efficiency, edge detection, Artificial neural network, business.industry, datasets, [INFO.INFO-MM]Computer Science [cs]/Multimedia [cs.MM], 020207 software engineering, Pattern recognition, neural networks, Computer Graphics and Computer-Aided Design, [INFO.INFO-GR]Computer Science [cs]/Graphics [cs.GR], Graphics (cs.GR), 020201 artificial intelligence & image processing, Artificial intelligence, Analysis tools, business, Efficient energy use
Abstract: In recent years, Convolutional Neural Networks (CNN) have proven to be efficient analysis tools for processing point clouds, e.g., for reconstruction, segmentation, and classification. In this article, we focus on the classification of edges in point clouds, where both edges and their surrounding are described. We propose a new parameterization adding to each point a set of differential information on its surrounding shape reconstructed at different scales. These parameters, stored in a Scale-Space Matrix (SSM) , provide a well-suited information from which an adequate neural network can learn the description of edges and use it to efficiently detect them in acquired point clouds. After successfully applying a multi-scale CNN on SSMs for the efficient classification of edges and their neighborhood, we propose a new lightweight neural network architecture outperforming the CNN in learning time, processing time, and classification capabilities. Our architecture is compact, requires small learning sets, is very fast to train, and classifies millions of points in seconds.
Published: 2021
Full Text: View/download PDF

3. Continual Self-Supervised Domain Adaptation for End-to-End Speaker Diarization

Author: Coria, Juan Manuel, Bredin, Hervé, Ghannay, Sahar, Rosset, Sophie, Laboratoire Interdisciplinaire des Sciences du Numérique (LISN), Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT), Centre National de la Recherche Scientifique (CNRS), Université Paris-Saclay under PhD contract number 2019-089, access to the HPC resources of IDRIS under the allocation AD011012177R1 made by GENCI, and IEEE Speech and Language Processing Technical Committee
Subjects: Self-supervised learning, Domain adaptation, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, End-to-end speaker diarization, Continual learning, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract: International audience; In conventional domain adaptation for speaker diarization, a large collection of annotated conversations from the target domain is required. In this work, we propose a novel continual training scheme for domain adaptation of an end-to-end speaker diarization system, which processes one conversation at a time and benefits from full self-supervision thanks to pseudo-labels. The qualities of our method allow for autonomous adaptation (e.g. of a voice assistant to a new household), while also avoiding permanent storage of possibly sensitive user conversations. We experiment extensively on the 11 domains of the DIHARD III corpus and show the effectiveness of our approach with respect to a pre-trained baseline, achieving a relative 17% performance improvement. We also find that data augmentation and a well-defined target domain are key factors to avoid divergence and to benefit from transfer.
Published: 2023
Full Text: View/download PDF

4. Analyse macroscopique des variations et modulations de F0 en lecture dans la maladie de Parkinson : données sur 320 locuteurs

Author: Vaysse, Robin, Ghio, Alain, Astésano, Corine, Farinas, Jérôme, Viallet, François, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Laboratoire Parole et Langage (LPL), Aix Marseille Université (AMU)-Centre National de la Recherche Scientifique (CNRS), Unité de Recherche Interdisciplinaire Octogone-Lordat (Octogone-Lordat), Université Toulouse - Jean Jaurès (UT2J), Praxiling (Praxiling), Université Paul-Valéry - Montpellier 3 (UPVM)-Centre National de la Recherche Scientifique (CNRS), AFCP : Association Francophone de la Communication Parlée, and ANR-18-CE45-0008,RUGBI,Recherche d'unités linguistiques pertinentes pour améliorer la mesure de l'intelligibilité de la parole altérée par des troubles de production pathologique(2018)
Subjects: speech, maladie de Parkinson, clinical phonetics, phonétique clinique, Parkinson disease, dysphonia, parole, dysarthria, L-dopa, dysarthrie, [INFO]Computer Science [cs], fundamental frequency, [SHS.LANGUE]Humanities and Social Sciences/Linguistics, dysphonie, fréquence fondamentale
Abstract: International audience; People with Parkinson’s disease (PD) usually suffer from laryngeal disorders involving fundamental frequency (F0) abnormalities. Despite the large number of studies of F0 on PD patients, no global consensus exists regarding the effects of this disease. This study aims to clarify these effects using a large corpora of read speech containing 320 speakers. The results show an increase in F0 for PD men compared to controls. A decrease in F0 variations is also observed for the patients regardless of gender. We also studied the effect of dopaminergic treatment inducing a slight increase in F0 and reducing the loss of F0 modulations in severely affected patients. Patient age appears to have the same impact on F0 in PD and healthy individuals. Our results are in partial agreement with pathophysiological models and contradict in some aspects studies done on smaller corpora.; Les personnes atteintes de la maladie de Parkinson (MDP) souffrent généralement de troubles d’origine laryngée impliquant des anomalies de fréquence fondamentale (F0). Malgré le nombre important d’études de la F0 chez les patients MDP, aucun consensus global n’existe concernant les effets de cette maladie. Cette étude cherche à clarifier ces effets grâce à un large corpus de parole lue (320 locuteurs). Les résultats montrent une élévation de F0 chez les hommes MDP. Nous observons également une réduction des modulations de F0 chez les patients indépendamment du sexe. L’effet du traitement dopaminergique induit une légère augmentation de F0 et réduit la perte des modulations de F0 chez les patients sévèrement atteints. L’âge des patients semble avoir le même impact sur la F0 des personnes MDP et saines. Nos résultats sont en accord partiel avec les modèles physiopathologiques et contredisent sur certains aspects des études faites sur des corpora plus petits.
Published: 2022
Full Text: View/download PDF

5. Audio-video fusion strategies for active speaker detection in meetings

Author: Lionel Pibre, Francisco Madrigal, Cyrille Equoy, Frédéric Lerasle, Thomas Pellegrini, Julien Pinquier, Isabelle Ferrané, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT), Équipe Robotique, Action et Perception (LAAS-RAP), Laboratoire d'analyse et d'architecture des systèmes (LAAS), Université de Toulouse (UT)-Université de Toulouse (UT)-Institut National des Sciences Appliquées - Toulouse (INSA Toulouse), Institut National des Sciences Appliquées (INSA)-Université de Toulouse (UT)-Institut National des Sciences Appliquées (INSA)-Université Toulouse - Jean Jaurès (UT2J), Projet LinTo - BPI France, and IRIT-LAAS / LinTo
Subjects: FOS: Computer and information sciences, Computer Science - Machine Learning, Sound (cs.SD), Multimodal fusion, Computer Networks and Communications, Computer Vision and Pattern Recognition (cs.CV), Computer Science - Computer Vision and Pattern Recognition, [INFO.INFO-MM]Computer Science [cs]/Multimedia [cs.MM], Audio processing, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Deep learning, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], Computer Science - Sound, Active speaker detection, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], Machine Learning (cs.LG), Video processing, Hardware and Architecture, Audio and Speech Processing (eess.AS), Media Technology, FOS: Electrical engineering, electronic engineering, information engineering, Speech analysis, Software, Electrical Engineering and Systems Science - Audio and Speech Processing
Abstract: International audience; Meetings are a common activity in professional contexts, and it remains challenging to endow vocal assistants with advanced functionalities to facilitate meeting management. In this context, a task like active speaker detection can provide useful insights to model interaction between meeting participants. Detection of the active speaker can be performed using only video based on the movements of the participants of a meeting. Depending on the assistant design and each participant position regarding the device, active speaker detection can benefit from information coming from visual and audio modalities. Motivated by our application context related to advanced meeting assistant, we want to combine audio and visual information to achieve the best possible performance. In this paper, we propose two different types of fusion (naive fusion and attention-based fusion) for the detection of the active speaker, combining two visual modalities and an audio modality through neural networks. In addition, the audio modality is mainly processed using neural networks. For comparison purpose, classical unsupervised approaches for audio feature extraction are also used. We expect visual data centered on the face of each participant to be very appropriate for detecting voice activity, based on the detection of lip and facial gestures. Thus, our baseline system uses visual data (video) and we chose a 3D Convolutional Neural Network (CNN) architecture, which is effective for simultaneously encoding appearance and movement. To improve this system, we supplemented the visual information by processing the audio stream with a CNN or an unsupervised speaker diarization system. We have further improved this system by adding visual modality information using motion through optical flow. We evaluated our proposal with a public and state-of-the-art benchmark: the AMI corpus. We analysed thecontribution of each system to the merger carried out in order to determine if a given participant is currently speaking. We also discussed the results we obtained. Besides, we have shown that, for our application context, adding motion information greatly improves performance. Finally, we have shown that attention-based fusion improves performance while reducing the standard deviation.
Published: 2022
Full Text: View/download PDF

6. Effects of oral and oropharyngeal cancer on speech intelligibility using acoustic analysis: Systematic review

Author: Julien Pinquier, Renée Speyer, Timothy Pommée, Mathieu Balaguer, Virginie Woisard, Jérôme Farinas, Centre National de la Recherche Scientifique - CNRS (FRANCE), Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), Centre Hospitalier Universitaire de Toulouse - CHU Toulouse (FRANCE), University of Oslo (NORWAY), Institut de Recherche en Informatique de Toulouse - IRIT (Toulouse, France), Centre Hospitalier Universitaire de Toulouse (CHU Toulouse), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT), Unité de Recherche Interdisciplinaire Octogone-Lordat (Octogone-Lordat), Université de Toulouse (UT)-Université de Toulouse (UT), Faculty of Educational Sciences [Oslo], University of Oslo (UiO), Agence Nationale de la Recherche, Grant/Award Number: ANR‐18‐CE45‐0008, Hospitals of Toulouse, PNRIA, ANR-18-CE45-0008,RUGBI,Recherche d'unités linguistiques pertinentes pour améliorer la mesure de l'intelligibilité de la parole altérée par des troubles de production pathologique(2018), European Project: 766287, CHU Toulouse [Toulouse], Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), and Université Fédérale Toulouse Midi-Pyrénées
Subjects: media_common.quotation_subject, Speech recognition, [SDV.CAN]Life Sciences [q-bio]/Cancer, Functional impact, Intelligibility (communication), behavioral disciplines and activities, Speech Acoustics, Machine Learning, 03 medical and health sciences, 0302 clinical medicine, Speech Production Measurement, Vowel, Perception, otorhinolaryngologic diseases, medicine, Humans, Speech, Head and neck cancer, speech, intelligibility, acoustics, head and neck cancer, oncology, Cancer, media_common, [SPI.ACOU]Engineering Sciences [physics]/Acoustics [physics.class-ph], Spectral moments, business.industry, Speech Intelligibility, Informatique et langage, Acoustics, 030206 dentistry, 3. Good health, Oropharyngeal Neoplasms, Systematic review, Oncology, Otorhinolaryngology, 030220 oncology & carcinogenesis, Speech disorder, medicine.symptom, Nasalance, business, Intelligibility, psychological phenomena and processes
Abstract: Background The development of automatic tools based on acoustic analysis allows to overcome the limitations of perceptual assessment for patients with head and neck cancer. The aim of this study is to provide a systematic review of literature describing the effects of oral and oropharyngeal cancer on speech intelligibility using acoustic analysis. Methods Two databases (PubMed and Embase) were surveyed. The selection process, according to the preferred reporting items for systematic reviews and meta‐analyses (PRISMA) statement, led to a final set of 22 articles. Results Nasalance is studied mainly in oropharyngeal patients. The vowels are mostly studied using formant analysis and vowel space area, the consonants by means of spectral moments with specific parameters according to their phonetic characteristic. Machine learning methods allow classifying “intelligible” or “unintelligible” speech for T3 or T4 tumors. Conclusions The development of comprehensive models combining different acoustic measures would allow a better consideration of the functional impact of the speech disorder.
Published: 2019
Full Text: View/download PDF

7. Weakly supervised discourse segmentation for multiparty oral conversations

Author: Gravellier, Lila, Hunter, Julie, Muller, Philippe, Pellegrini, Thomas, Ferrané, Isabelle, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Linagora Labs [Toulouse], Linagora [Puteaux], MEthodes et ingénierie des Langues, des Ontologies et du DIscours (IRIT-MELODI), LinTO (projet PIA/BPI France No P169201), ACL: Association for Computational Linguistics, ANR-20-CE23-0017,SUMM-RE,Supervision Distante pour les Compte-Rendus Enrichis de Relations Rhétoriques(2020), ANR-19-P3IA-0004,ANITI,Artificial and Natural Intelligence Toulouse Institute(2019), ANR-18-CE23-0005,LUDAU,Découverte peu et non-supervisée d'unités audio à l'aide d'apprentissage profond(2018), Hunter, Julie, Supervision Distante pour les Compte-Rendus Enrichis de Relations Rhétoriques - - SUMM-RE2020 - ANR-20-CE23-0017 - AAPG2020 - VALID, Artificial and Natural Intelligence Toulouse Institute - - ANITI2019 - ANR-19-P3IA-0004 - P3IA - VALID, and APPEL À PROJETS GÉNÉRIQUE 2018 - Découverte peu et non-supervisée d'unités audio à l'aide d'apprentissage profond - - LUDAU2018 - ANR-18-CE23-0005 - AAPG2018 - VALID
Subjects: [INFO.INFO-CL] Computer Science [cs]/Computation and Language [cs.CL], 0202 electrical engineering, electronic engineering, information engineering, 020201 artificial intelligence & image processing, 02 engineering and technology, 010501 environmental sciences, 16. Peace & justice, 01 natural sciences, [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], 0105 earth and related environmental sciences
Abstract: International audience; Discourse segmentation, the first step of discourse analysis, has been shown to improve results for text summarization, translation and other NLP tasks. While segmentation models for written text tend to perform well, they are not directly applicable to spontaneous, oral conversation, which has linguistic features foreign to written text. Segmentation is less studied for this type of language, where annotated data is scarce, and existing corpora more heterogeneous. We develop a weak supervision approach to adapt, using minimal annotation, a state of the art discourse segmenter trained on written text to French conversation transcripts. Supervision is given by a latent model bootstrapped by manually defined heuristic rules that use linguistic and acoustic information. The resulting model improves the original segmenter, especially in contexts where information on speaker turns is lacking or noisy, gaining up to 13% in F-score. Evaluation is performed on data like those used to define our heuristic rules, but also on transcripts from two other corpora.
Published: 2021
Full Text: View/download PDF

8. End-to-end speaker segmentation for overlap-aware resegmentation

Author: Antoine Laurent, Hervé Bredin, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Laboratoire d'Informatique de l'Université du Mans (LIUM), Le Mans Université (UM), HPC resources of IDRIS under the allocation AD011012177 made by GENCI (Grand Équipement National de Calcul Intensif), ANR-16-CE92-0025,PLUMCOT,Identification non-supervisée des personnages de films et séries télévisées(2016), and ANR-19-CE38-0012,GEM,Mesure de l'égalité entre les sexes dans les médias(2019)
Subjects: FOS: Computer and information sciences, speaker segmentation, Sound (cs.SD), Computer science, Speech recognition, overlapped speech detection, 02 engineering and technology, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], Computer Science - Sound, [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], Task (project management), [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], 030507 speech-language pathology & audiology, 03 medical and health sciences, End-to-end principle, Audio and Speech Processing (eess.AS), 0202 electrical engineering, electronic engineering, information engineering, FOS: Electrical engineering, electronic engineering, information engineering, Segmentation, resegmentation, Diarization error rate, Voice activity detection, 020206 networking & telecommunications, Speaker diarisation, voice activity detection, Temporal resolution, speaker diarization, 0305 other medical science, Change detection, Electrical Engineering and Systems Science - Audio and Speech Processing
Abstract: Speaker segmentation consists in partitioning a conversation between one or more speakers into speaker turns. Usually addressed as the late combination of three sub-tasks (voice activity detection, speaker change detection, and overlapped speech detection), we propose to train an end-to-end segmentation model that does it directly. Inspired by the original end-to-end neural speaker diarization approach (EEND), the task is modeled as a multi-label classification problem using permutation-invariant training. The main difference is that our model operates on short audio chunks (5 seconds) but at a much higher temporal resolution (every 16ms). Experiments on multiple speaker diarization datasets conclude that our model can be used with great success on both voice activity detection and overlapped speech detection. Our proposed model can also be used as a post-processing step, to detect and correctly assign overlapped speech regions. Relative diarization error rate improvement over the best considered baseline (VBx) reaches 17% on AMI, 13% on DIHARD 3, and 13% on VoxConverse., Comment: Camera-ready version for Interspeech 2021 with significantly better voice activity detection, overlapped speech detection, and speaker diarization results. The code used for results reported in v1 contained a small bug that has now been fixed
Published: 2021
Full Text: View/download PDF

9. Le Carcinologic Handicap Index (CHI) : questionnaire d’autoévaluation du handicap à destination des patients traités pour cancer des VADS

Author: Josiane Percodani, Mathieu Balaguer, Virginie Woisard, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, CHU Toulouse [Toulouse], Centre National de la Recherche Scientifique - CNRS (FRANCE), Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), Centre Hospitalier Universitaire de Toulouse - CHU Toulouse (FRANCE), and Institut National Polytechnique de Toulouse - INPT (FRANCE)
Subjects: [INFO.INFO-DB]Computer Science [cs]/Databases [cs.DB], Cancer de la tête et du cou, Qualité de vie, Autoévaluation, Handicap, [INFO.INFO-RO]Computer Science [cs]/Operations Research [cs.RO], 3. Good health, 03 medical and health sciences, 0302 clinical medicine, Otorhinolaryngology, Base de données, 030211 gastroenterology & hepatology, Surgery, 030223 otorhinolaryngology, Recherche opérationnelle
Abstract: National audience; Contexte : 5ème cancer en termes d'incidence en France, le cancer des voies aérodigestives supérieures (VADS) impacte de façon importante la vie quotidienne des sujets atteints, tant en termes d'activités que de participation. Il devient donc nécessaire d'évaluer au préalable les dimensions fonctionnelles et psychosociales : nous avons ainsi développé le Carcinologic Handicap Index (ou CHI). Objectif(s) : Cette étude prospective a pour objet l'élaboration et la validation d'un questionnaire d'autoévaluation du handicap après traitement des cancers des VADS. Méthode : Le CHI a été conçu de façon empirique sur le modèle du « Voice handicap index ». Il comprend 9 dimensions autoévaluées par le patient (douleur, déglutition, nutrition, respiration, phonation, audition, vision, olfaction-gustation, psychosocial). Chacune comporte 4 questions à coter en termes de fréquence, ce qui permet le calcul d'un score pour chaque dimension ainsi qu'un score global. Il a été testé auprès de 86 patients constituant la population pathologique (majorité d'hommes, âge moyen de 59 ans, présentant une tumeur des voies aérodigestives supérieures) et 18 sujets (population témoin), pour validation du questionnaire. Résultats : La cohérence interne globale est à 0,905 (test alpha de Cronbach). La validité de contenu (r, entre les scores du questionnaire et les EVA correspondantes) se situe entre 0,6 et 0,8 sauf pour le domaine Vision et pour le score total par rapport à l'EVA de la santé générale (r ¿ 0,5), la pathologie n'altérant pas directement la vision et n'étant qu'un élément parmi d'autre pouvant jouer sur la santé générale). La validité temporelle est satisfaisante (r > 0,7 avec p < 0,0001) sauf pour le domaine Respiration (r = 0,624, pouvant s'expliquer par la fluctuation de l'encombrement pulmonaire). Selon le site de la lésion, les domaines Déglutition, Nutrition et Respiration sont impactés différemment. Enfin, nous ne retrouvons pas de différence significative entre patients et témoins pour les domaines Douleur, Audition et Vision. Discussion : Le CHI présente des qualités psychométriques acceptables et peut être considéré comme un véritable outil clinique destiné aux professionnels de santé permettant de prendre en compte l'impact de la pathologie ORL sur la qualité de vie des patients, principalement concernant les domaines fonctionnels directement atteints par le cancer ou son traitement. http://congresccfuo.fr/
Published: 2017
Full Text: View/download PDF

10. Low-activity supervised convolutional spiking neural networks applied to speech commands recognition

Author: Thomas Pellegrini, Timothée Masquelier, Romain Zimmer, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse Capitole (UT Capitole), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse Capitole (UT Capitole), Université Fédérale Toulouse Midi-Pyrénées, Centre de recherche cerveau et cognition (CERCO UMR5549), Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Toulouse Mind & Brain Institut (TMBI), Calcul en Midi-Pyrénées, France - CALMIP (Grant 2020-p20022), ANR-18-CE23-0005,LUDAU,Découverte peu et non-supervisée d'unités audio à l'aide d'apprentissage profond(2018), ANR-19-P3IA-0004,ANITI,Artificial and Natural Intelligence Toulouse Institute(2019), Pellegrini, Thomas, APPEL À PROJETS GÉNÉRIQUE 2018 - Découverte peu et non-supervisée d'unités audio à l'aide d'apprentissage profond - - LUDAU2018 - ANR-18-CE23-0005 - AAPG2018 - VALID, Artificial and Natural Intelligence Toulouse Institute - - ANITI2019 - ANR-19-P3IA-0004 - P3IA - VALID, Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Centre de recherche cerveau et cognition (CERCO), Institut des sciences du cerveau de Toulouse. (ISCT), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-CHU Toulouse [Toulouse]-Institut National de la Santé et de la Recherche Médicale (INSERM)-Centre National de la Recherche Scientifique (CNRS)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), and Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-CHU Toulouse [Toulouse]-Institut National de la Santé et de la Recherche Médicale (INSERM)-Centre National de la Recherche Scientifique (CNRS)-Centre National de la Recherche Scientifique (CNRS)
Subjects: FOS: Computer and information sciences, Computer Science - Machine Learning, Sound (cs.SD), surrogate gradient, Computer science, Word error rate, Biological neuron model, 02 engineering and technology, Regularization (mathematics), [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], Computer Science - Sound, Convolution, Machine Learning (cs.LG), Audio and Speech Processing (eess.AS), 0202 electrical engineering, electronic engineering, information engineering, FOS: Electrical engineering, electronic engineering, information engineering, speech command recognition, Spiking neural network, Computer Science - Computation and Language, Spiking neural networks, business.industry, 020206 networking & telecommunications, Pattern recognition, Speech processing, Backpropagation, [INFO.INFO-SD] Computer Science [cs]/Sound [cs.SD], Task (computing), [INFO.INFO-CL] Computer Science [cs]/Computation and Language [cs.CL], [INFO.INFO-SD]Computer Science [cs]/Sound [cs.SD], 020201 artificial intelligence & image processing, Artificial intelligence, business, Computation and Language (cs.CL), Electrical Engineering and Systems Science - Audio and Speech Processing
Abstract: Deep Neural Networks (DNNs) are the current state-of-the-art models in many speech related tasks. There is a growing interest, though, for more biologically realistic, hardware friendly and energy efficient models, named Spiking Neural Networks (SNNs). Recently, it has been shown that SNNs can be trained efficiently, in a supervised manner, using backpropagation with a surrogate gradient trick. In this work, we report speech command (SC) recognition experiments using supervised SNNs. We explored the Leaky-Integrate-Fire (LIF) neuron model for this task, and show that a model comprised of stacked dilated convolution spiking layers can reach an error rate very close to standard DNNs on the Google SC v1 dataset: 5.5%, while keeping a very sparse spiking activity, below 5%, thank to a new regularization term. We also show that modeling the leakage of the neuron membrane potential is useful, since the LIF model outperformed its non-leaky model counterpart significantly., Comment: Accepted to IEEE Spoken Language Technology Workshop 2021
Published: 2020
Full Text: View/download PDF

11. The Airbus Air Traffic Control Speech Recognition 2018 Challenge: Towards ATC Automatic Transcription and Call Sign Detection

Author: Estelle Delpech, François Lancelot, Thomas Pellegrini, Jérôme Farinas, Airbus (FRANCE), Centre National de la Recherche Scientifique - CNRS (FRANCE), Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Airbus [France], and Institut National Polytechnique de Toulouse - INPT (FRANCE)
Subjects: FOS: Computer and information sciences, Specialized language, Sign detection, Sound (cs.SD), 050210 logistics & transportation, Computer science, Speech recognition, 05 social sciences, Word error rate, Informatique et langage, 02 engineering and technology, Air traffic control, Speech processing, [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], Computer Science - Sound, Transcription (linguistics), Audio and Speech Processing (eess.AS), 13. Climate action, 0502 economics and business, FOS: Electrical engineering, electronic engineering, information engineering, 0202 electrical engineering, electronic engineering, information engineering, 020201 artificial intelligence & image processing, Electrical Engineering and Systems Science - Audio and Speech Processing
Abstract: In this paper, we describe the outcomes of the challenge organized and run by Airbus and partners in 2018. The challenge consisted of two tasks applied to Air Traffic Control (ATC) speech in English: 1) automatic speech-to-text transcription, 2) call sign detection (CSD). The registered participants were provided with 40 hours of speech along with manual transcriptions. Twenty-two teams submitted predictions on a five hour evaluation set. ATC speech processing is challenging for several reasons: high speech rate, foreign-accented speech with a great diversity of accents, noisy communication channels. The best ranked team achieved a 7.62% Word Error Rate and a 82.41% CSD F1-score. Transcribing pilots' speech was found to be twice as harder as controllers' speech. Remaining issues towards solving ATC ASR are also discussed., 5 pages, 4 tables, 1 figure
Published: 2019
Full Text: View/download PDF

12. The Acoustic Voice Quality Index, Version 03.01, in French and the Voice Handicap Index

Author: Camille Finck, Timothy Pommée, Youri Maryn, Dominique Morsomme, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Université de Liège, University of Antwerp (UA), Universiteit Gent = Ghent University [Belgium] (UGENT), University College Ghent, Centre Hospitalier Universitaire de Liège (CHU-Liège), and European Union's Horizon 2020 research and innovation programme under Marie Skłodowska-Curie grant agreement No 766287
Subjects: Adult, Male, Quality of life, medicine.medical_specialty, Index (economics), Voice Quality, Voice assessment, media_common.quotation_subject, Acoustic voice analysis, Audiology, Severity of Illness Index, Spearman's rank correlation coefficient, Correlation, Disability Evaluation, Young Adult, 030507 speech-language pathology & audiology, 03 medical and health sciences, Speech and Hearing, 0302 clinical medicine, Cost of Illness, Phonation, Speech Production Measurement, Predictive Value of Tests, Surveys and Questionnaires, otorhinolaryngologic diseases, Quantitative assessment, medicine, Humans, [INFO]Computer Science [cs], Quality (business), Voice Handicap Index, 030223 otorhinolaryngology, Aged, media_common, Aged, 80 and over, Reproducibility of Results, Acoustic voice quality index, Acoustics, Middle Aged, Dysphonia, LPN and LVN, University hospital, Otorhinolaryngology, Female, Human medicine, Voice handicap index, 0305 other medical science, Psychology
Abstract: International audience; Objectives. The Acoustic Voice Quality Index (AVQI) version 03.01 is a tool for quantitative assessment of the overall severity of dysphonia. Its computation includes six acoustic parameters, which are all carried out by the freeware Praat. It is based on the recordings of a sustained vowel and a part of a text read aloud. The psychometric qualities of this tool have been confirmed by numerous publications in various languages, including French. However, studies investigating the correlation between tools for objective vocal assessment and voice-related quality of life show inconsistent results. Hence, the aim of this study was to contribute to the debate in measuring the correlations between the AVQI 03.01 score computed on French samples and the Voice Handicap Index. Methods. The data of 78 patients was used, collected during initial vocal assessment and stored in the ENT caseload of the University Hospital of Liège. The authors measured the spearman rank-order correlation between the VHI total and sub-scores and the scores computed by the AVQI 03.01 on French samples. Eventually, the correlation between the diagnostic decisions (“normophonia” vs “dysphonia”) of both tools was measured using Cramer's phi. Results. The Spearman correlation between the AVQI 03.01 score and the total VHI score, when controlling for age, is moderate (rs = .62, p < .0001). The correlations between the AVQI 03.01 score and the functional, emotional and physical sub-scores of the VHI are similarly moderate (rs = .643, .543 and .514, respectively, p
Published: 2020
Full Text: View/download PDF

13. Improving Vehicle Re-Identification using CNN Latent Spaces: Metrics Comparison and Track-to-track Extension

Author: André Péninou, Alain Crouzil, Vincent Charvillat, Patrice Guyot, Sylvie Chambon, Julien Pinquier, Geoffrey Roman-Jimenez, Thierry Malon, Florence Sèdes, Christine Sénac, Systèmes d’Informations Généralisées (IRIT-SIG), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, SolutionData Group : Architecte de données (SolutionData), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), EuroMov - Digital Health in Motion (Euromov DHM), IMT - MINES ALES (IMT - MINES ALES), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Montpellier (UM), CoMputational imagINg anD viSion (IRIT-MINDS), Real Expression Artificial Life (IRIT-REVA), Institut National Polytechnique (Toulouse) (Toulouse INP), Université Toulouse III - Paul Sabatier (UT3), IUT Toulouse 2 Blagnac, and Université Toulouse - Jean Jaurès (UT2J)
Subjects: FOS: Computer and information sciences, Computer Science - Machine Learning, Computer science, Computer Vision and Pattern Recognition (cs.CV), Computer applications to medicine. Medical informatics, Computer Science - Computer Vision and Pattern Recognition, R858-859.7, 02 engineering and technology, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE], Track (rail transport), Measure (mathematics), Re identification, Machine Learning (cs.LG), QA76.75-76.765, image-to-track, 0202 electrical engineering, electronic engineering, information engineering, Computer software, business.industry, track-to-track, Process (computing), 020206 networking & telecommunications, Pattern recognition, Extension (predicate logic), Euclidean distance, [INFO.INFO-IR]Computer Science [cs]/Information Retrieval [cs.IR], [INFO.INFO-TI]Computer Science [cs]/Image Processing [eess.IV], Metric (mathematics), 020201 artificial intelligence & image processing, Computer Vision and Pattern Recognition, Artificial intelligence, business, Neural coding, CNN latent spaces, vehicle re‐identification, Software
Abstract: This paper addresses the problem of vehicle re-identification using distance comparison of images in CNN latent spaces. Firstly, we study the impact of the distance metrics, comparing performances obtained with different metrics: the minimal Euclidean distance (MED), the minimal cosine distance (MCD), and the residue of the sparse coding reconstruction (RSCR). These metrics are applied using features extracted from five different CNN architectures, namely ResNet18, AlexNet, VGG16, InceptionV3 and DenseNet201. We use the specific vehicle re-identification dataset VeRi to fine-tune these CNNs and evaluate results. In overall, independently of the CNN used, MCD outperforms MED, commonly used in the literature. These results are confirmed on other vehicle retrieval datasets. Secondly, we extend the state-of-the-art image-to-track process (I2TP) to a track-to-track process (T2TP). The three distance metrics are extended to measure distance between tracks, enabling T2TP. We compared T2TP with I2TP using the same CNN models. Results show that T2TP outperforms I2TP for MCD and RSCR. T2TP combining DenseNet201 and MCD-based metrics exhibits the best performances, outperforming the state-of-the-art I2TP-based models. Finally, experiments highlight two main results: i) the impact of metric choice in vehicle re-identification, and ii) T2TP improves the performances compared to I2TP, especially when coupled with MCD-based metrics., Comment: This paper is a postprint of a paper submitted to and accepted for publication in the journal IET Computer Vision and is subject to Institution of Engineering and Technology Copyright. The copy of record is available at the IET Digital Library
Published: 2019
Full Text: View/download PDF

14. A multi-modal perception based architecture for a non-intrusive domestic assistant robot

Author: Christophe Mollaret, Alhayat Ali Mekonnen, Julien Pinquier, Frederic Lerasle, Isabelle Ferrane, Équipe Robotique, Action et Perception (LAAS-RAP), Laboratoire d'analyse et d'architecture des systèmes (LAAS), Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Institut National des Sciences Appliquées - Toulouse (INSA Toulouse), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Toulouse III - Paul Sabatier (UT3), ACM, IEEE, ANR-12-CORD-0003,RIDDLE,Robots perceptuels et interactifs dédiés aux environnements quotidiens(2012), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Institut National des Sciences Appliquées - Toulouse (INSA Toulouse), Institut National des Sciences Appliquées (INSA)-Université de Toulouse (UT)-Institut National des Sciences Appliquées (INSA)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), and Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3)
Subjects: 0209 industrial biotechnology, 05 social sciences, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Detectors, 02 engineering and technology, Robot kinematics, Speech recognition, Magnetic heads, 020901 industrial engineering & automation, Speech, [INFO.INFO-RB]Computer Science [cs]/Robotics [cs.RO], 0501 psychology and cognitive sciences, Head, 050107 human factors
Abstract: International audience; We present a multi-modal perception based architecture to realize a non-intrusive domestic assistant robot. The realized robot is non-intrusive in that it only starts interaction with a user when it detects the user's intention to do so automatically. All the robot's actions are based on multi-modal perceptions, which include: user detection based on RGB-D data, user's intention-for-interaction detection with RGB-D and audio data, and communication via speech recognition. The utilization of multi-modal cues in different parts of the robotic activity paves the way to successful robotic runs.
Published: 2016
Full Text: View/download PDF

15. Twist your logic with TouIST

Author: Ben Slimane, Skander, Comte, Alexis, Gasquet, Olivier, Heba, Abdelawad, Lezaud, Olivier, Maris, Frédéric, Valais, Maël, Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées, Logique, Interaction, Langue et Calcul (IRIT-LILaC), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Argumentation, Décision, Raisonnement, Incertitude et Apprentissage (IRIT-ADRIA), Institut National Polytechnique de Toulouse - INPT (FRANCE), Centre National de la Recherche Scientifique - CNRS (FRANCE), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), and Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE)
Subjects: Logique en informatique, FOS: Computer and information sciences, Problem solving, Computer Science - Artificial Intelligence, K.3.2, I.2.8, SAT-prover, Search, [INFO.INFO-LO]Computer Science [cs]/Logic in Computer Science [cs.LO], Informatique et langage, Intelligence artificielle, Interface, Apprentissage, [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], High-level logic formalization, Computer Science - Computers and Society, Control Methods, TheoryofComputation_MATHEMATICALLOGICANDFORMALLANGUAGES, Artificial Intelligence (cs.AI), [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Computers and Society (cs.CY)
Abstract: SAT provers are powerful tools for solving real-sized logic problems, but using them requires solid programming knowledge and may be seen w.r.t.\ logic like assembly language w.r.t.\ programming. Something like a high level language was missing to ease various users to take benefit of these tools. {\sc \texttt {TouIST}}\ aims at filling this gap. It is devoted to propositional logic and its main features are 1) to offer a high-level logic langage for expressing succintly complex formulas (e.g.\ formulas describing Sudoku rules, planification problems,\ldots) and 2) to find models to these formulas by using the adequate powerful prover, which the user has no need to know about. It consists in a friendly interface that offers several syntactic facilities and which is connected with some sufficiently powerful provers allowing to automatically solve big instances of difficult problems (such as time-tables or Sudokus). It can interact with various provers: pure SAT solver but also SMT provers (SAT modulo theories - like linear theory of reals, etc) and thus may also be used by beginners for experiencing with pure propositional problems up to graduate students or even researchers for solving planification problems involving big sets of fluents and numerical constraints on them., Comment: Proceedings of the Fourth International Conference on Tools for Teaching Logic (TTL2015), Rennes, France, June 9-12, 2015. Editors: M. Antonia Huertas, Jo\~ao Marcos, Mar\'ia Manzano, Sophie Pinchinat, Fran\c{c}ois Schwarzentruber
Published: 2015
Full Text: View/download PDF

16. Hierarchical Hidden Markov Model in Detecting Activities of Daily Living in Wearable Videos for Studies of Dementia

Author: Jenny Benois-Pineau, Svebor Karaman, Régine André-Obrecht, Rémi Mégret, Yann Gaëstel, Julien Pinquier, Vladislavs Dovgalecs, Jean-François Dartigues, Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Laboratoire de l'intégration, du matériau au système (IMS), Université Sciences et Technologies - Bordeaux 1-Institut Polytechnique de Bordeaux-Centre National de la Recherche Scientifique (CNRS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Université Toulouse III - Paul Sabatier (UT3), Institut de Santé Publique, d'Epidémiologie et de Développement (ISPED), Université Bordeaux Segalen - Bordeaux 2, Epidémiologie et Biostatistique [Bordeaux], Université Bordeaux Segalen - Bordeaux 2-Institut de Santé Publique, d'Épidémiologie et de Développement (ISPED)-Institut National de la Santé et de la Recherche Médicale (INSERM), and IMMED
Subjects: FOS: Computer and information sciences, Computer Networks and Communications, Computer science, ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION, Wearable computer, Context (language use), Hidden Markov Model, Motion (physics), Task (project management), Activities of Daily Living, Media Technology, Computer vision, Segmentation, Hidden Markov model, business.industry, Search engine indexing, [INFO.INFO-MM]Computer Science [cs]/Multimedia [cs.MM], Hierarchical hidden Markov model, Wearable Videos, Multimedia (cs.MM), Hardware and Architecture, Artificial intelligence, Video Indexing, business, Software, Computer Science - Multimedia
Abstract: International audience; This paper presents a method for indexing activities of daily living in videos obtained from wearable cameras. In the context of dementia diagnosis by doctors, the videos are recorded at patients' houses and later visualized by the medical practitioners. The videos may last up to two hours, therefore a tool for an efficient navigation in terms of activities of interest is crucial for the doctors. The specific recording mode provides video data which are really difficult, being a single sequence shot where strong motion and sharp lighting changes often appear. Our work introduces an automatic motion based segmentation of the video and a video structuring approach in terms of activities by a hierarchical two-level Hidden Markov Model. We define our description space over motion and visual characteristics of video and audio channels. Experiments on real data obtained from the recording at home of several patients show the difficulty of the task and the promising results of our approach.
Published: 2011
Full Text: View/download PDF

17. Activities of Daily Living Indexing by Hierarchical HMM for Dementia Diagnostics

Author: Jenny Benois-Pineau, Svebor Karaman, Rémi Mégret, Jean-François Dartigues, Yann Gaëstel, Julien Pinquier, Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Epidémiologie et Biostatistique [Bordeaux], Université Bordeaux Segalen - Bordeaux 2-Institut de Santé Publique, d'Épidémiologie et de Développement (ISPED)-Institut National de la Santé et de la Recherche Médicale (INSERM), Institut de Santé Publique, d'Epidémiologie et de Développement (ISPED), Université Bordeaux Segalen - Bordeaux 2, Laboratoire de l'intégration, du matériau au système (IMS), Université Sciences et Technologies - Bordeaux 1-Institut Polytechnique de Bordeaux-Centre National de la Recherche Scientifique (CNRS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, ANR Blanc, IEEE, and IMMED
Subjects: FOS: Computer and information sciences, Activities of daily living, Computer science, Early signs, ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION, Wearable computer, 02 engineering and technology, Aged Dementia, 0202 electrical engineering, electronic engineering, information engineering, medicine, Dementia, Computer vision, Hidden Markov model, Hidden Markov Models, Activities of Daily Living Indexing, business.industry, Search engine indexing, [INFO.INFO-MM]Computer Science [cs]/Multimedia [cs.MM], 020207 software engineering, medicine.disease, Multimedia (cs.MM), 020201 artificial intelligence & image processing, Artificial intelligence, Video Indexing, business, Merge (version control), Computer Science - Multimedia
Abstract: This paper presents a method for indexing human ac- tivities in videos captured from a wearable camera being worn by patients, for studies of progression of the dementia diseases. Our method aims to produce indexes to facilitate the navigation throughout the individual video recordings, which could help doctors search for early signs of the dis- ease in the activities of daily living. The recorded videos have strong motion and sharp lighting changes, inducing noise for the analysis. The proposed approach is based on a two steps analysis. First, we propose a new approach to segment this type of video, based on apparent motion. Each segment is characterized by two original motion de- scriptors, as well as color, and audio descriptors. Second, a Hidden-Markov Model formulation is used to merge the multimodal audio and video features, and classify the test segments. Experiments show the good properties of the ap- proach on real data., Comment: 2011 9th International Workshop on Content-Based Multimedia Indexing (CBMI), Madrid : Spain (2011)
Published: 2011
Full Text: View/download PDF

18. Face-and-clothing based people clustering in video content

Author: Christine Sénac, Philippe Joly, Elie Khoury, Laboratoire d'Informatique de l'Université du Mans (LIUM), Le Mans Université (UM), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), and Université Fédérale Toulouse Midi-Pyrénées
Subjects: Matching (statistics), Computer science, business.industry, Correlation clustering, Search engine indexing, face, ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION, 020207 software engineering, Pattern recognition, 02 engineering and technology, Hierarchical clustering, clothing, video people indexing, Histogram, Face (geometry), 0202 electrical engineering, electronic engineering, information engineering, Canopy clustering algorithm, [INFO]Computer Science [cs], 020201 artificial intelligence & image processing, Computer vision, Artificial intelligence, Cluster analysis, business, hierarchical clustering
Abstract: International audience; Content-based people clustering is a crucial step for people indexing within video documents. In this paper, we investigate the use of both face and clothing features. A method of extracting a keyface for each video sequence is proposed. An algorithm based on the average of the N-minimum pair distances between local invariant features is used in order to resolve the problem of face matching. An original method for clothing matching is proposed based on 3D histogram of the dominant color. A 3-levels hierarchical bottom-up clustering that combines local invariant features, skin color, 3D histogram and clothing texture is also described. Experiments and results show the efficiency of the proposed clustering system.
Published: 2010
Full Text: View/download PDF

19. Application of the EMD Decomposition to Discriminate Nasalized vs. Vowels Phones in French

Author: Régine André-Obrecht, Olivier Pietquin, Mohamed Saïdi, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, SUPELEC-Campus Metz, Ecole Supérieure d'Electricité - SUPELEC (FRANCE), and Université Toulouse III - Paul Sabatier (UT3)
Subjects: Artificial neural network, Computer science, Speech recognition, Feature extraction, Mode (statistics), Nasalized phones, Signal, Hilbert–Huang transform, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, MFCC, EMD, Preprocessor, Mel-frequency cepstrum, Vowels, ANN, IMF, [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing, Nasal vowel
Abstract: International audience; This work deals with the application of the Empirical Mode Decomposition (EMD) with the goal of showing its capabilities and limitations when applied to nasalized vs. oral vowels phones classification. The method used in this study consists in three classical stages: signal preprocessing, feature extraction and decision. Firstly, the speech signal is decomposed using the EMD method so as to extract the three first Intrinsic Mode Functions (IMF). Then, Mel-Frequency Cepstral Coefficients (MFCC) are extracted from these IMFs or a (maybe partial) sum of them. Finally, an Artificial Neural Network (ANN) is used to distinguish nasal vowels from oral vowels in French (French database Bref80). Besides the fact that this study resulted in a significant improvement in the level of discrimination, when we use our method compared to the standard application of MFCC to the original signal. It has also allowed us to know which IMFs allows to better characterize the nasal vowels from the oral vowels.
Published: 2010
Full Text: View/download PDF

20. A novel strategy for speaker verification based on SVM classification of pairs of speech sequences

Author: K. Daoudi, J. Louradour, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), and Université Fédérale Toulouse Midi-Pyrénées
Subjects: Speaker verification, Computer science, business.industry, Speech recognition, 020206 networking & telecommunications, Pattern recognition, 02 engineering and technology, Speaker recognition, Speaker diarisation, Support vector machine, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, Bayesian information criterion, 0202 electrical engineering, electronic engineering, information engineering, NIST, 020201 artificial intelligence & image processing, Segmentation, Artificial intelligence, business, [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing, Classifier (UML), ComputingMilieux_MISCELLANEOUS
Abstract: We introduce a novel strategy for speaker verification based on the conception of a classifier which is independent of the target speaker, as opposed to traditional systems where the classifier is always target dependent. The basic principle is to build a system that decides whether two sequences were pronounced by the same speaker. In our view, this system is aimed to complement traditional ones. We borrow techniques from the speaker segmentation area, namely the Bayesian Information Criterion (BIC), to conceive a kernel between pairs of sequences. We then use this kernel to implement our new system in an SVM scheme. We present experiments on NIST SRE data using the Biosecure project protocol. The individual performance of the new system is poor as compared to the baseline UBM-GMM and the GLDS-SVM. However, as expected, the fusion leads to better performances.
Published: 2007
Full Text: View/download PDF

21. Modeling prosody for language identification on read and spontaneous speech

Author: François Pellegrino, Jean-Luc Rouas, Régine André-Obrecht, Jérôme Farinas, Laboratorio de Sistemas de Lingua Falada (L2F), Instituto de Engenharia de Sistemas e Computadores (INESC), Laboratoire Bordelais de Recherche en Informatique (LaBRI), Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université Sciences et Technologies - Bordeaux 1-Université Bordeaux Segalen - Bordeaux 2, Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI), Université Paris-Sud - Paris 11 (UP11)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Saclay (COmUE), Laboratoire Electronique, Ondes et Signaux pour les Transports (INRETS/LEOST), Institut National de Recherche sur les Transports et leur Sécurité (INRETS), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées, Dynamique Du Langage (DDL), Université Lumière - Lyon 2 (UL2)-Centre National de la Recherche Scientifique (CNRS), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Université Paris Saclay (COmUE)-Centre National de la Recherche Scientifique (CNRS)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Université Paris-Sud - Paris 11 (UP11), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), and Université Toulouse III - Paul Sabatier (UT3)
Subjects: Language identification, Computer science, Speech recognition, 02 engineering and technology, computer.software_genre, 030507 speech-language pathology & audiology, 03 medical and health sciences, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, Stress (linguistics), 0202 electrical engineering, electronic engineering, information engineering, SAMOVA ART.ps IAL, Prosody, ComputingMilieux_MISCELLANEOUS, Phonotactics, business.industry, 020206 networking & telecommunications, Speech corpus, 16. Peace & justice, Speech processing, Identification (information), Artificial intelligence, 0305 other medical science, business, computer, Natural language processing, Natural language
Abstract: This paper deals with an approach to automatic language identification using only prosodic modeling. The actual approach for language identification focuses mainly on phonotactics because it gives the best results. We propose here to evaluate the relevance of prosodic information for language identification with read studio recording, as in our previous experiment (Rouas, J.L. et al., Proc. Eusipeo'02, 2002), and spontaneous telephone speech. For read speech, experiments were performed on the five languages of the MULTEXT database (Campione, E. and Veronis, J., Proc. ICSLP'98, p.3163-6, 1998). On the MULTEXT corpus, our prosodic system achieved an identification rate of 79 % on the five languages discrimination task. For spontaneous speech, experiments are made on the ten languages of the OGI multilingual telephone speech corpus (Muthusamy, Y.K. et al., Proc. ICSLP'92, 1992). On the OGI MLTS corpus, the results are given for language pair discrimination tasks, and are compared with results from F. Cummins et al. (Technical Report idsia-07-99, Istituto Dalle Molle di Studi sull'Intelligenza Artificiale, Lugano, CH, 1999). In conclusion, if our prosodic system achieves good performance on read speech, it might not take into account the complexity of spontaneous speech prosody.
Published: 2003
Full Text: View/download PDF

22. IRIM at TRECVID 2010: High Level Feature Extraction and Instance Search

Author: David Gorisse, Frédéric Precioso, Philippe Gosselin, Lionel Granjon, denis pellerin, Michèle Rombaut, Hervé Bredin, Lionel Koenig, Hélène Lachambre, Elie El Khoury, Rémi Vieux, Boris Mansencal, Yifan Zhou, Jenny Benois-Pineau, Hervé Jégou, Stéphane Ayache, Bahjat Safadi, Yubing Tong, Franck Thollard, Georges Quénot, Alexandre Benoît, Patrick Lambert, Equipes Traitement de l'Information et Systèmes (ETIS - UMR 8051), CY Cergy Paris Université (CY)-Centre National de la Recherche Scientifique (CNRS)-Ecole Nationale Supérieure de l'Electronique et de ses Applications (ENSEA), CRIL Ingénierie [Rennes], CRIL Ingénierie, GIPSA-Services (GIPSA-Services), Grenoble Images Parole Signal Automatique (GIPSA-lab), Université Pierre Mendès France - Grenoble 2 (UPMF)-Université Stendhal - Grenoble 3-Université Joseph Fourier - Grenoble 1 (UJF)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Pierre Mendès France - Grenoble 2 (UPMF)-Université Stendhal - Grenoble 3-Université Joseph Fourier - Grenoble 1 (UJF)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS), GIPSA - Architecture, Géométrie, Perception, Images, Gestes (GIPSA-AGPIG), Département Images et Signal (GIPSA-DIS), Université Pierre Mendès France - Grenoble 2 (UPMF)-Université Stendhal - Grenoble 3-Université Joseph Fourier - Grenoble 1 (UJF)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Université Pierre Mendès France - Grenoble 2 (UPMF)-Université Stendhal - Grenoble 3-Université Joseph Fourier - Grenoble 1 (UJF)-Institut polytechnique de Grenoble - Grenoble Institute of Technology (Grenoble INP )-Centre National de la Recherche Scientifique (CNRS)-Grenoble Images Parole Signal Automatique (GIPSA-lab), Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI), Université Paris Saclay (COmUE)-Centre National de la Recherche Scientifique (CNRS)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Université Paris-Sud - Paris 11 (UP11), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Multimedia content-based indexing (TEXMEX), Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Rennes 1 (UR1), Université de Rennes (UNIV-RENNES)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées (INSA)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS)-Université de Rennes 1 (UR1), Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées (INSA)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS)-Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria), Laboratoire d'informatique Fondamentale de Marseille - UMR 6166 (LIF), Université de la Méditerranée - Aix-Marseille 2-Université de Provence - Aix-Marseille 1-Centre National de la Recherche Scientifique (CNRS), Laboratoire d'Informatique de Grenoble (LIG), 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), Laboratoire Hubert Curien [Saint Etienne] (LHC), Institut d'Optique Graduate School (IOGS)-Université Jean Monnet [Saint-Étienne] (UJM)-Centre National de la Recherche Scientifique (CNRS), Modélisation et Recherche d’Information Multimédia [Grenoble] (MRIM), 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)-Université Pierre Mendès France - Grenoble 2 (UPMF)-Université Joseph Fourier - Grenoble 1 (UJF), Laboratoire d'Informatique, Systèmes, Traitement de l'Information et de la Connaissance (LISTIC), Université Savoie Mont Blanc (USMB [Université de Savoie] [Université de Chambéry]), Ecole Nationale Supérieure de l'Electronique et de ses Applications (ENSEA)-Centre National de la Recherche Scientifique (CNRS)-CY Cergy Paris Université (CY), Université Paris-Sud - Paris 11 (UP11)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Saclay (COmUE), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT), Université de Bordeaux (UB)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Centre National de la Recherche Scientifique (CNRS), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre National de la Recherche Scientifique (CNRS)-Inria Rennes – Bretagne Atlantique, 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), Laboratoire Hubert Curien (LHC), Institut d'Optique Graduate School (IOGS)-Université Jean Monnet - Saint-Étienne (UJM)-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)-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), and Fauvet, Marie-Christine
Subjects: [INFO.INFO-IR]Computer Science [cs]/Information Retrieval [cs.IR], InformationSystems_INFORMATIONSTORAGEANDRETRIEVAL, ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION, [INFO.INFO-IR] Computer Science [cs]/Information Retrieval [cs.IR]
Abstract: International audience; The IRIM group is a consortium of French teams working on Multimedia Indexing and Retrieval. This paper describes our participation to the TRECVID 2010 semantic indexing and instance search tasks. For the semantic indexing task, we evaluated a number of different descriptors and tried different fusion strategies, in particular hierarchical fusion. The best IRIM run has a Mean Inferred Average Precision of 0.0442, which is above the task median performance. We found that fusion of the classification scores from different classifier types improves the performance and that even with a quite low individual performance, audio descriptors can help. For the instance search task, we used only one of the example images in our queries. The rank is nearly in the middle of the list of participants. The experiment showed that HSV features outperform the concatenation of HSV and edge histograms or the wavelet feature.

23. Automatic Modelling of Rhythm and Intonation for Language Identification

Author: Jean-Luc ROUAS, Jérôme Farinas, François Pellegrino, Laboratorio de Sistemas de Lingua Falada (L2F), Instituto de Engenharia de Sistemas e Computadores (INESC), Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI), Université Paris Saclay (COmUE)-Centre National de la Recherche Scientifique (CNRS)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Université Paris-Sud - Paris 11 (UP11), Laboratoire Electronique, Ondes et Signaux pour les Transports (INRETS/LEOST), Institut National de Recherche sur les Transports et leur Sécurité (INRETS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Dynamique Du Langage (DDL), Université Lumière - Lyon 2 (UL2)-Centre National de la Recherche Scientifique (CNRS), Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université Sciences et Technologies - Bordeaux 1-Université Bordeaux Segalen - Bordeaux 2, Université Paris-Sud - Paris 11 (UP11)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Saclay (COmUE), Université Toulouse 1 Capitole (UT1)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), and Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP)
Subjects: [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, SAMOVA ART.ps IAL, ComputingMilieux_MISCELLANEOUS
Abstract: International audience; no abstract

24. Audio-Video Event Recognition System For Public Transport Security

Author: Sébastien Ambellouis, Nicolas Allezard, Jean-Luc Rouas, Van-Thinh Vu, Amaury Flancquart, Quoc Cuong Pham, Monique Thonnat, Patrick Sayd, Francois Bremond, Gabriele Davini, Intelligent environments for the resolution of problems for autonomous systems (ORION), Inria Sophia Antipolis - Méditerranée (CRISAM), Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria), Département d'Electronique, des Détecteurs et d'Informatique (ex SEDI) (DEDI), Institut de Recherches sur les lois Fondamentales de l'Univers (IRFU), Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Université Paris-Saclay-Commissariat à l'énergie atomique et aux énergies alternatives (CEA)-Université Paris-Saclay, Laboratorio de Sistemas de Lingua Falada (L2F), Instituto de Engenharia de Sistemas e Computadores (INESC), Laboratoire Bordelais de Recherche en Informatique (LaBRI), Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université Sciences et Technologies - Bordeaux 1-Université Bordeaux Segalen - Bordeaux 2, Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI), Université Paris-Sud - Paris 11 (UP11)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Saclay (COmUE), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées, Laboratoire Electronique, Ondes et Signaux pour les Transports (INRETS/LEOST), Institut National de Recherche sur les Transports et leur Sécurité (INRETS), Département d'Electronique, des Détecteurs et d'Informatique pour la Physique (ex SEDI) (DEDIP), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Université Paris Saclay (COmUE)-Centre National de la Recherche Scientifique (CNRS)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Université Paris-Sud - Paris 11 (UP11), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université de Bordeaux (UB)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Centre National de la Recherche Scientifique (CNRS), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), and Université de Toulouse (UT)
Subjects: Event ontology, Event (computing), business.industry, Computer science, Event recognition, ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION, 02 engineering and technology, computer.software_genre, Tracking (particle physics), Facial recognition system, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, 020204 information systems, Public transport, 0202 electrical engineering, electronic engineering, information engineering, 020201 artificial intelligence & image processing, Computer vision, Artificial intelligence, Audio signal processing, business, Face detection, computer, [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing
Abstract: International audience; This paper presents an audio-video surveillance system for the automatic surveillance in public transport vehicle. The system comprises six modules including in particular three novel ones: (i) Face Detection and Tracking, (ii) Audio Event Detection and (iii) Audio-Video Scenario Recognition. The Face Detection and Tracking module is responsible for detecting and tracking faces of people in front of cameras. The Audio Event Detection module detects abnormal audio events which are precursor for detecting scenarios which have been predefined by end-users. The Audio-Video Scenario Recognition module performs high level interpretation of the observed objects by combining audio and video events based on spatio-temporal reasoning. The performance of the system is evaluated for a series of pre-defined audio, video and audio-video events specified using an audio-video event ontology.

25. Détection de la parole et de la musique : fusion de deux approches

Author: Julien Pinquier, Julie Mauclair, Jean-Luc ROUAS, Régine André-Obrecht, Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées, Laboratoire d'Informatique de l'Université du Mans (LIUM), Le Mans Université (UM), Laboratorio de Sistemas de Lingua Falada (L2F), Instituto de Engenharia de Sistemas e Computadores (INESC), Laboratoire Bordelais de Recherche en Informatique (LaBRI), Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université Sciences et Technologies - Bordeaux 1-Université Bordeaux Segalen - Bordeaux 2, Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI), Université Paris-Sud - Paris 11 (UP11)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Saclay (COmUE), Laboratoire Electronique, Ondes et Signaux pour les Transports (INRETS/LEOST), Institut National de Recherche sur les Transports et leur Sécurité (INRETS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Université Paris Saclay (COmUE)-Centre National de la Recherche Scientifique (CNRS)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Université Paris-Sud - Paris 11 (UP11), and Université Toulouse III - Paul Sabatier (UT3)
Subjects: [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, ComputingMilieux_MISCELLANEOUS, SAMOVA ART.ps indexation
Abstract: International audience; no abstract

26. Soundscape visualization: A new approach based on automatic annotation and Samocharts

Author: Guyot, P., Julien PINQUIER, Institut National Polytechnique de Toulouse - INPT (FRANCE), Centre National de la Recherche Scientifique - CNRS (FRANCE), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, and Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE)
Subjects: Traitement des images, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, [INFO.INFO-TI]Computer Science [cs]/Image Processing [eess.IV], [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Traitement du signal et de l'image, Vision par ordinateur et reconnaissance de formes, Intelligence artificielle, Synthèse d'image et réalité virtuelle, Soundscape visualization, [INFO.INFO-GR]Computer Science [cs]/Graphics [cs.GR], [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]
Abstract: International audience; The visualization of sounds facilitates their identification and classification. However, in the case of audio recording websites, the access to a sound is usually based on the metadata of the sounds, i.e.sources and recording conditions. As sonic environments, or soundscapes, are mostly composed of multiples sources, their compact description is an issue that makes difficult the choice of an item in a sound corpus. The time-component matrix chart, which is abbreviated as TM-chart, has been proposed recently as a tool to describe and compare sonic environments. However their process of creation is based on a subjective annotation that makes their creation time-consuming. In this paper, we present a new method for urban soundscape corpus visualization. In the context of the CIESS project, we propose amochart: an extension of the TM-chart that is based on sound detection algorithms. We describe three original algorithms that allow the detection of alarms, footsteps, and motors. Samocharts can be computed from the results of these algorithms. This process is applied to a concrete case study: 20 urban recordings of 5 minutes each, from different situations (places and time). An application case shows that Samocharts allow an identification of different situations. Finally, the whole method provides a low-cost tool for soundscape visualization that can easily be applied to the management and use of a sound corpus.

27. A new measure to predict the a priori performance of automatic transcription systems on reverberated speech

Author: Sébastien Ferreira, Julien Pinquier, Jérôme Farinas, Julie Mauclair, Stéphane Rabant, Authôt (FRANCE), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), and Université Fédérale Toulouse Midi-Pyrénées
Subjects: [INFO]Computer Science [cs]
Abstract: International audience; Advances in the field of Automatic Speech Recognition (ASR) make it possible to use this technology in less and less controlled environments. However, reverberation remains a challenge. The purpose of this study is to predict the a priori reverberation impact on the ASR performance. We analysed the statistical behaviour of the vocal excitation of the vowels and created a measure based on this observation. This measure is called Excitation Behaviour (EB). To find the link between the EB and the Word Error Rate (WER) obtained by the transcription system, a regression model was calculated. To evaluate the performance of the regression model we observed the mean prediction error. We also used the same protocol on two other measures: Speech-to-Reverberation Modulation energy Ratio (SRMR) and Spectral Decay Distribution (SDD). The speech corpus used is the Wall Street Journal (WSJ0 and WSJ1) corpus. Speech was artificially reverberated using Room Impulse Response (RIR) from the REVERB challenge corpus. Recorded RIRs allow to simulate 7 different reverberation conditions: 3 rooms with different sizes (Small, Medium and Large) with 2 types of distances between a speaker and a microphone array (near=50cm and far=200 cm), and the last condition is without reverberation. ASR system was trained with the subset trainₛi284, the regression model was trained with dev93 and eval92 was used to test the prediction. It shows that the EB obtains an average prediction error of 13.88 while the SRMR obtains 17.87 and the SDD 17.48. Using 20 utterances (approximately 2m20s), the average prediction error decreases to 7.63 for the EB, 13.08 for the SRMR and 13.02 for the SDD. EB measure is better correlated to ASR performance than other reverberation measures.

28. Automatic analysis of word association data from the Evolex psycholinguistic tasks using computational lexical semantic similarity measures

Author: Gaume, Bruno, Tanguy, Ludovic, Fabre, Cécile, Ho-Dac, Lydia-Mai, Pierrejean, Bénédicte, Hathout, Nabil, Jérôme Farinas, Pinquier, Julien, Danet, Lola, Péran, Patrice, Boissezon, Xavier, Jucla, Mélanie, Centre National de la Recherche Scientifique - CNRS (FRANCE), Institut National Polytechnique de Toulouse - INPT (FRANCE), Institut National de la Santé et de la Recherche Médicale - INSERM (FRANCE), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), Toulouse Neuro Imaging Center - TONIC (Toulouse, France), Cognition, Langues, Langage, Ergonomie (CLLE-ERSS), Université Bordeaux Montaigne-École pratique des hautes études (EPHE), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-Université Toulouse - Jean Jaurès (UT2J)-Centre National de la Recherche Scientifique (CNRS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Toulouse Neuro Imaging Center (ToNIC), Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Institut National de la Santé et de la Recherche Médicale (INSERM)-Hôpital Purpan [Toulouse], CHU Toulouse [Toulouse]-CHU Toulouse [Toulouse], Service de Médecine Physique et Réadaptation, CHU Toulouse [Toulouse]-Hôpital de Rangueil, CHU Toulouse [Toulouse], Octogone Unité de Recherche Interdisciplinaire (Octogone), Université Toulouse - Jean Jaurès (UT2J), and Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE)
Subjects: [SDV.NEU.SC]Life Sciences [q-bio]/Neurons and Cognition [q-bio.NC]/Cognitive Sciences, Informatique et langage, [SCCO.LING]Cognitive science/Linguistics, [INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL], Word association data
Abstract: International audience; This paper is the fruit of a multidisciplinary project gathering researchers in Psycholinguistics, Neuropsychology, Computer Science, Natural Language Processing and Linguistics. It proposes a new data-based inductive method for automatically characterising the relation between pairs of words collected in psycholinguistics experiments on lexical access. This method takes advantage of four complementary computational measures of semantic similarity. We compare these techniques by assessing their correlation with a manual categorisation of 559 distinct word pairs, and with the distribution of data produced by 30 test subjects. We show that some measures are more correlated than others with the frequency of lexical associations, and that they also differ in the way they capture different semantic relations. This allows us to consider building a multidimensional lexical similarity to automate the classification of lexical associations.

29. On using generic profiles and views for dynamic web services adaptation

Author: Sébastien Laborie, Bouchra Soukkarieh, Florence Sèdes, Institut National Polytechnique de Toulouse - INPT (FRANCE), Centre National de la Recherche Scientifique - CNRS (FRANCE), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), Systèmes d’Informations Généralisées (IRIT-SIG), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Laboratoire Informatique de l'Université de Pau et des Pays de l'Adour (LIUPPA), Université de Pau et des Pays de l'Adour (UPPA), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), and Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE)
Subjects: Théorie de l'information, [INFO.INFO-IT]Computer Science [cs]/Information Theory [cs.IT], [INFO.INFO-IR]Computer Science [cs]/Information Retrieval [cs.IR], Recherche d'information, Web Service, Context-Aware adaptation
Abstract: International audience; The emergence of mobile technologies allows users to be connected to services anytime, anywhere and anyhow. However, this mobility requires new constraints on the execution of the services and their presentations to users. Hence, these ones must be selected and adapted according to not only the user's profile but also his context. In this paper, we present a new context-aware adaptation architecture based on Web Services, named CA-WIS. Our proposal aims users to interact with web services that correspond to their needs and their contexts.

30. Audio Events Detection in Public Transport Vehicle

Author: Jérôme Louradour, Sébastien Ambellouis, Jean-Luc Rouas, Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Laboratorio de Sistemas de Lingua Falada (L2F), Instituto de Engenharia de Sistemas e Computadores (INESC), Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI), Université Paris Saclay (COmUE)-Centre National de la Recherche Scientifique (CNRS)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Université Paris-Sud - Paris 11 (UP11), Laboratoire Electronique, Ondes et Signaux pour les Transports (INRETS/LEOST), Institut National de Recherche sur les Transports et leur Sécurité (INRETS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université Sciences et Technologies - Bordeaux 1-Université Bordeaux Segalen - Bordeaux 2, Université Paris-Sud - Paris 11 (UP11)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Saclay (COmUE), Université Toulouse 1 Capitole (UT1)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), and Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP)
Subjects: Engineering, business.industry, Speech recognition, Search engine indexing, 020206 networking & telecommunications, 02 engineering and technology, computer.software_genre, Control room, Support vector machine, symbols.namesake, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, Public transport, 11. Sustainability, Audio analyzer, 0202 electrical engineering, electronic engineering, information engineering, symbols, 020201 artificial intelligence & image processing, Railway engineering, Data mining, Audio signal processing, business, Gaussian process, computer, [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing
Abstract: International audience; This paper addresses the problem of automatic audio analysis for aided surveillance application in public transport. The aim of such application is to detect critical situations and to warn the control room. We propose a comparative study of two methods of modelisation/classification of acoustical segments. The problem is quite similar to the "'audio indexing"' framework, nevertheless the environment here is very noisy. We present two general frameworks based on Gaussian Model Mixture (GMM) and Support Vector Machine (SVM) to achieve shout detection in railway embedded environment.

31. Using automatic speech recognition for the prediction of impaired speech identification

Author: Lionel Fontan, Imed Laaridh, Jérôme Farinas, Julien Pinquier, Maxime Le Coz, Christian Füllgrabe, Archean Labs, Laboratoire Informatique d'Avignon (LIA), Avignon Université (AU)-Centre d'Enseignement et de Recherche en Informatique - CERI, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Loughborough University, Project: CarlW, and Grélaud, Françoise
Subjects: [INFO]Computer Science [cs], [INFO] Computer Science [cs], Automatic Speech Recognition (ASR)
Abstract: International audience; Age-related hearing loss (ARHL) is a very prevalent hearing disorder in adults that negatively impacts on the ability to understand speech, especially in noisy environments. The most common rehabilitation strategy is to fit hearing aids (HAs). Their benefit is generally assessed by measuring speech-identification performance with and without HAs. However, such so-called “speech audiometry” can be fairly lengthy, and its results are likely to be influenced by the patient's level of fatigue, cognitive state and familiarity with the speech material used for the assessment. In order to overcome these issues, the feasibility of using objective measures based on automatic speech recognition (ASR) to predict human speech-identification performances was recently investigated (Fontan et al., 2017; Fontan et al., in preparation; Kollmeier et al., 2016). Here, we present the results of a series of experiments, that combined ASR and an ARHL simulation to predict human performances for various tasks ranging from phoneme discrimination to sentences identification. More specifically, signal processing techniques (Nejime & Moore, 1997) were used to process the speech tokens to mimic some of the perceptual consequences of ARHL on speech perception (i.e., elevated thresholds, reduced frequency selectivity and loudness recruitment), and the processed speech tokens were then fed to an ASR system for analysis. To provide “proof-of-concept”, our first experiments focussed on the prediction of unaided speech perception in quiet, while subsequent experiments investigated the applicability of the ASR system to aided and unaided speech perception in noise.

32. Automatic Speech Recognition Predicts Speech Intelligibility and Comprehension for Listeners With Simulated Age-Related Hearing Loss

Author: Lionel Fontan, Xavier Aumont, Cynthia Magnen, Julien Pinquier, Julien Tardieu, Jérôme Farinas, Pascal Gaillard, Christian Füllgrabe, Isabelle Ferrané, Centre National de la Recherche Scientifique - CNRS (FRANCE), Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), University of Nottingham (UNITED KINGDOM), Archean Technologies (FRANCE), Institut de Recherche en Informatique de Toulouse - IRIT (Toulouse, France), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Maison des Sciences de l'Homme et de la Société de Toulouse (MSHS-T), Centre National de la Recherche Scientifique (CNRS)-Université Fédérale Toulouse Midi-Pyrénées, Cognition, Langues, Langage, Ergonomie (CLLE), École pratique des hautes études (EPHE), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-Université Toulouse - Jean Jaurès (UT2J)-Centre National de la Recherche Scientifique (CNRS), Medical Research Council Institute of Hearing Research, School of Medicine, The University of Nottingham, Nottinghamshire, Projet AGIL-IT, région Midi-Pyrénées, Centre National de la Recherche Scientifique (CNRS)-École pratique des hautes études (EPHE), and Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-Université Toulouse - Jean Jaurès (UT2J)
Subjects: Male, Aging, genetic structures, Speech recognition, Presbycusis, Intelligibility (communication), Language and Linguistics, Pattern Recognition, Automated, Speech Recognition Software, Traitement des images, 0302 clinical medicine, Hearing Aids, Traitement du signal et de l'image, [SHS.LANGUE]Humanities and Social Sciences/Linguistics, 030223 otorhinolaryngology, Child, Amplification or hearing aids, Speech perception, Automatic speech recognition, Vision par ordinateur et reconnaissance de formes, Middle Aged, [SCCO.PSYC]Cognitive science/Psychology, Regression Analysis, Female, medicine.symptom, Psychology, Comprehension, Auditory perception, Adult, Linguistics and Language, Adolescent, Hearing loss, [SCCO.COMP]Cognitive science/Computer science, Assessment, Speech comprehension, 03 medical and health sciences, Speech and Hearing, Young Adult, medicine, otorhinolaryngologic diseases, Humans, Computer Simulation, Hearing Loss, Synthèse d'image et réalité virtuelle, Speech Intelligibility, Intelligence artificielle, medicine.disease, Speech processing, 030217 neurology & neurosurgery
Abstract: PurposeThe purpose of this article is to assess speech processing for listeners with simulated age-related hearing loss (ARHL) and to investigate whether the observed performance can be replicated using an automatic speech recognition (ASR) system. The long-term goal of this research is to develop a system that will assist audiologists/hearing-aid dispensers in the fine-tuning of hearing aids.MethodSixty young participants with normal hearing listened to speech materials mimicking the perceptual consequences of ARHL at different levels of severity. Two intelligibility tests (repetition of words and sentences) and 1 comprehension test (responding to oral commands by moving virtual objects) were administered. Several language models were developed and used by the ASR system in order to fit human performances.ResultsStrong significant positive correlations were observed between human and ASR scores, with coefficients up to .99. However, the spectral smearing used to simulate losses in frequency selectivity caused larger declines in ASR performance than in human performance.ConclusionBoth intelligibility and comprehension scores for listeners with simulated ARHL are highly correlated with the performances of an ASR-based system. In the future, it needs to be determined if the ASR system is similarly successful in predicting speech processing in noise and by older people with ARHL.
Full Text: View/download PDF

33. Application de l'analyse factorielle discriminante à l'identification des langues

Author: Jorge Gutiérrez, Jean-Luc ROUAS, Régine André-Obrecht, Equipes Traitement de l'Information et Systèmes (ETIS - UMR 8051), Ecole Nationale Supérieure de l'Electronique et de ses Applications (ENSEA)-Université de Cergy Pontoise (UCP), Université Paris-Seine-Université Paris-Seine-Centre National de la Recherche Scientifique (CNRS), Equipe Circuit Instrumentation et Modélisation Electronique (ECIME), Université de Cergy Pontoise (UCP), Université Paris-Seine-Université Paris-Seine-Ecole Nationale Supérieure de l'Electronique et de ses Applications, STMicroelectronics [Crolles] (ST-CROLLES), Laboratorio de Sistemas de Lingua Falada (L2F), Instituto de Engenharia de Sistemas e Computadores (INESC), Laboratoire Bordelais de Recherche en Informatique (LaBRI), Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université Sciences et Technologies - Bordeaux 1-Université Bordeaux Segalen - Bordeaux 2, Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI), Université Paris-Sud - Paris 11 (UP11)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Saclay (COmUE), Laboratoire Electronique, Ondes et Signaux pour les Transports (INRETS/LEOST), Institut National de Recherche sur les Transports et leur Sécurité (INRETS), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées, Rouas, Jean-Luc, CY Cergy Paris Université (CY)-Centre National de la Recherche Scientifique (CNRS)-Ecole Nationale Supérieure de l'Electronique et de ses Applications (ENSEA), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Université Paris Saclay (COmUE)-Centre National de la Recherche Scientifique (CNRS)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Université Paris-Sud - Paris 11 (UP11), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), and Université Toulouse III - Paul Sabatier (UT3)
Subjects: [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, [INFO.INFO-TS] Computer Science [cs]/Signal and Image Processing, ComputingMilieux_MISCELLANEOUS
Abstract: International audience; no abstract

34. Rhythmic unit extraction and modelling for automatic language identification

Author: François Pellegrino, Régine André-Obrecht, Jérôme Farinas, Jean-Luc Rouas, Laboratorio de Sistemas de Lingua Falada (L2F), Instituto de Engenharia de Sistemas e Computadores (INESC), Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI), Université Paris Saclay (COmUE)-Centre National de la Recherche Scientifique (CNRS)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Université Paris-Sud - Paris 11 (UP11), Laboratoire Electronique, Ondes et Signaux pour les Transports (INRETS/LEOST), Institut National de Recherche sur les Transports et leur Sécurité (INRETS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Dynamique Du Langage (DDL), Université Lumière - Lyon 2 (UL2)-Centre National de la Recherche Scientifique (CNRS), Université Toulouse III - Paul Sabatier (UT3), Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université Sciences et Technologies - Bordeaux 1-Université Bordeaux Segalen - Bordeaux 2, Université Paris-Sud - Paris 11 (UP11)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Saclay (COmUE), Université Toulouse 1 Capitole (UT1)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), and Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP)
Subjects: Linguistics and Language, Language identification, Computer science, Speech recognition, Mandarin Chinese, 050105 experimental psychology, Language and Linguistics, 030507 speech-language pathology & audiology, 03 medical and health sciences, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, Vowel, 0501 psychology and cognitive sciences, Phonotactics, Communication, 05 social sciences, Phonetics, language.human_language, Computer Science Applications, Identification (information), Duration (music), Modeling and Simulation, language, Computer Vision and Pattern Recognition, Syllable, 0305 other medical science, [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing, Software
Abstract: International audience; This paper deals with an approach to Automatic Language Identification based on rhythmic modelling. Beside phonetics and phonotactics, rhythm is actually one of the most promising features to be considered for language identification, even if its extraction and modelling are not a straightforward issue. Actually, one of the main problems to address is what to model. In this paper, an algorithm of rhythm extraction is described: using a vowel detection algorithm, rhythmic units related to syllables are segmented. Several parameters are extracted (consonantal and vowel duration, cluster complexity) and modelled with a Gaussian Mixture. Experiments are performed on read speech for 7 languages (English, French, German, Italian, Japanese, Mandarin and Spanish) and results reach up to 86 ± 6% of correct discrimination between stress-timed mora-timed and syllable-timed classes of languages, and to 67 ± 8% percent of correct language identification on average for the 7 languages with utterances of 21 seconds. These results are commented and compared with those obtained with a standard acoustic Gaussian mixture modelling approach (88 ± 5% of correct identification for the 7-languages identification task).

35. Automatic Estimation of Speaking Rate in Multilingual Spontaneous Speech

Author: François Pellegrino, Jérôme Farinas, Jean-Luc Rouas, Dynamique Du Langage (DDL), Université Lumière - Lyon 2 (UL2)-Centre National de la Recherche Scientifique (CNRS), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées, Laboratorio de Sistemas de Lingua Falada (L2F), Instituto de Engenharia de Sistemas e Computadores (INESC), Laboratoire Bordelais de Recherche en Informatique (LaBRI), Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université Sciences et Technologies - Bordeaux 1-Université Bordeaux Segalen - Bordeaux 2, Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI), Université Paris-Sud - Paris 11 (UP11)-Sorbonne Université - UFR d'Ingénierie (UFR 919), Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris Saclay (COmUE), Laboratoire Electronique, Ondes et Signaux pour les Transports (INRETS/LEOST), Institut National de Recherche sur les Transports et leur Sécurité (INRETS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Université Paris Saclay (COmUE)-Centre National de la Recherche Scientifique (CNRS)-Sorbonne Université - UFR d'Ingénierie (UFR 919), and Sorbonne Université (SU)-Sorbonne Université (SU)-Université Paris-Saclay-Université Paris-Sud - Paris 11 (UP11)
Subjects: [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, SAMOVA ART.ps IAL, ComputingMilieux_MISCELLANEOUS
Abstract: International audience; no abstract

36. Bazinga! A Dataset for Multi-Party Dialogues Structuring

Author: Paul Lerner, Juliette Bergoënd, Camille Guinaudeau, Hervé Bredin, Benjamin Maurice, Sharleyne Lefevre, Martin Bouteiller, Aman Berhe, Léo Galmant, Ruiqing Yin, Claude Barras, Laboratoire Interdisciplinaire des Sciences du Numérique (LISN), Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS), Traitement du Langage Parlé (TLP ), Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Sciences et Technologies des Langues (STL), Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT), Centre National de la Recherche Scientifique (CNRS), Information, Langue Ecrite et Signée (ILES), Vocapia Research [Orsay], Vocapia, European Language Resources Association (ELRA), StoryArcs project (2016-1267D), ANR-16-CE92-0025,PLUMCOT,Identification non-supervisée des personnages de films et séries télévisées(2016), and ANR-19-CE23-0028,MEERQAT,Représentation multimédia d'entités et systèmes de question réponse(2019)
Subjects: Multi-Party Dialogues, [INFO.INFO-MM]Computer Science [cs]/Multimedia [cs.MM], Entity Linking, Speaker Diarization
Abstract: International audience; We introduce a dataset built around a large collection of TV (and movie) series. Those are filled with challenging multi-party dialogues. Moreover, TV series come with a very active fan base that allows the collection of metadata and accelerates annotation. With 16 TV and movie series, Bazinga! amounts to 400+ hours of speech and 8M+ tokens, including 500K+ tokens annotated with the speaker, addressee, and entity linking information. Along with the dataset, we also provide a baseline for speaker diarization, punctuation restoration, and person entity recognition. The results demonstrate the difficulty of the tasks and of transfer learning from models trained on mono-speaker audio or written text, which is more widely available. This work is a step towards better multi-party dialogue structuring and understanding. Bazinga! is available at hf.co/bazinga. Because (a large) part of Bazinga! is only partially annotated, we also expect this dataset to foster research towards self-or weakly-supervised learning methods.

37. Speaker age estimation for elderly speech recognition in European Portuguese

Author: Annika Hämäläinen, Vahid Hedayati, Miguel Sales Dias, Isabel Trancoso, Thomas Pellegrini, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT), Université Toulouse III - Paul Sabatier (UT3), Instituto de Engenharia de Sistemas e Computadores Investigação e Desenvolvimento em Lisboa (INESC-ID), Instituto Superior Técnico, Universidade Técnica de Lisboa (IST)-Instituto de Engenharia de Sistemas e Computadores (INESC), Universidade de Lisboa = University of Lisbon (ULISBOA), Instituto Universitário de Lisboa (ISCTE-IUL), FCT Fundacao para a Cîencia e a Tecnologia, under projects PTDC/EEA-PLP/121111/2010, PTDC/EIA-CCO/122542/2010, and PEst-OE/EEI/LA0021/2013, QREN 5329 Fala Global project, which is co-funded by Microsoft and the European Structural Funds for Portugal (FEDER) through POR Lisboa (Regional Operational Programme of Lisbon), as part of the National Strategic Reference Framework (QREN), the national program of incentives for Portuguese businesses and industry, International Speech Communication Association (ISCA), European Project: 611396,EC:FP7:ICT,FP7-ICT-2013-SME-DCA,SPEDIAL(2013), Chng E.S. , Li H., Meng H., Ma B. and Xie L, Centre National de la Recherche Scientifique - CNRS (FRANCE), Instituto de Engenharia de Sistemas e Computadores - Investigação e Desenvolvimento - INESC-ID (PORTUGAL), Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE), Instituto Universitario de Lisboa - ISCTE IUL (PORTUGAL), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), Microsoft (USA), and Universidade de Lisboa - ULisboa (PORTUGAL)
Subjects: Speech production, Computer science, Speech recognition, Automatic speech recognition, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Vision par ordinateur et reconnaissance de formes, Intelligence artificielle, Elderly speech, Ciências Naturais::Ciências Físicas [Domínio/Área Científica], I-vector extraction, [INFO.INFO-GR]Computer Science [cs]/Graphics [cs.GR], language.human_language, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], Traitement des images, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, European Portuguese, Transcription (linguistics), Age estimation, [INFO.INFO-TI]Computer Science [cs]/Image Processing [eess.IV], Automatic age estimation, language, Traitement du signal et de l'image, Synthèse d'image et réalité virtuelle
Abstract: Phone-like acoustic models (AMs) used in large-vocabulary automatic speech recognition (ASR) systems are usually trained with speech collected from young adult speakers. Using such models, ASR performance may decrease by about 10% absolute when transcribing elderly speech. Ageing is known to alter speech production in ways that require ASR systems to be adapted, in particular at the level of acoustic modeling. In this study, we investigated automatic age estimation in order to select age-specific adapted AMs. A large corpus of read speech from European Portuguese speakers aged 60 or over was used. Age estimation (AE) based on i-vectors and support vector regression achieved mean error rates of about 4.2 and 4.5 years for males and females, respectively. Compared with a baseline ASR system with AMs trained using young adult speech and a WER of 13.9%, the selection of five-year-range adapted AMs, based on the estimated age of the speakers, led to a decrease in WER of about 9.3% relative (1.3% absolute). Comparable gains in ASR performance were observed when considering two larger age ranges (60-75 and 76-90) instead of six five-year ranges, suggesting that it would be sufficient to use the two large ranges only. info:eu-repo/semantics/acceptedVersion

38. What do pause patterns in non-fluent aphasia tell us about monitoring speech? A study of morpho-syntactic complexity, accuracy and fluency in agrammatic sentence and connected discourse production

Author: Sahraoui, Halima, Mauclair, Julie, Lorraine BAQUÉ, Nespoulous, Jean-Luc, Centre National de la Recherche Scientifique - CNRS (FRANCE), Institut National Polytechnique de Toulouse - INPT (FRANCE), Université Toulouse III - Paul Sabatier - UT3 (FRANCE), Université Toulouse - Jean Jaurès - UT2J (FRANCE), Université Toulouse 1 Capitole - UT1 (FRANCE), Universitat Autònoma de Barcelona - UAB (SPAIN), Université Paris Descartes - Paris V (FRANCE), Unité de Recherche Interdisciplinaire Octogone-Lordat (Octogone-Lordat), Université Toulouse - Jean Jaurès (UT2J), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1)-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées, Université Paris Descartes - Paris 5 (UPD5), Universitat Autònoma de Barcelona [Barcelona] (UAB), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Universitat Autònoma de Barcelona (UAB), and Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE)
Subjects: Non fluent aphasia, lcsh:BF1-990, Agrammatism, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Vision par ordinateur et reconnaissance de formes, Intelligence artificielle, Speech monitoring, [INFO.INFO-GR]Computer Science [cs]/Graphics [cs.GR], Pause, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], 030507 speech-language pathology & audiology, 03 medical and health sciences, Traitement des images, 0302 clinical medicine, lcsh:Psychology, [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing, [INFO.INFO-TI]Computer Science [cs]/Image Processing [eess.IV], Fluency, Traitement du signal et de l'image, Adaptation, 0305 other medical science, Synthèse d'image et réalité virtuelle, 030217 neurology & neurosurgery, General Psychology
Abstract: Compared to normal speech, agrammatic utterances are “telegraphic” with frequent closed-class words omissions, simple clauses, short utterances and reduced morpho-syntactic complexity. Besides, its predominant symptom comprises fluency disturbances and great efforts needed to produce speech. Results from previous corpora analyses confirm that variability in agrammatic performance is a key feature for understanding impaired and strategic language use (Kolk & Heeschen, 1990 ; Kolk, 2006, Sahraoui & Nespoulous, 2012). Beyond the reduced on-line processing resource related to the underlying impairment, patients adjust the surface structure to be encoded according to contextual constraints. Moreover, across-task variability in fluency may also be determined by the use of various adaptation strategies, such as elliptical and corrective encoding strategies, related to the focus on form enabling better grammatical accuracy under certain conditions. In particular, corrective strategies are due to the preserved ability, in these patients, to detect errors and monitor their speech production (Postma, 2000 ; Oomen, Postma & Kolk, 2001). This suggests that agrammatic patients may over-use the monitoring device in producing language, at a pre-articulatory or post-articulatory stage and according to the type of task (Sahraoui, 2014). In this study, we carefully look at temporal aspects of agrammatic speech production in order to understand how far does speech (non)fluency and dysfluency show a relation to the morpho-syntactic properties of sentence and connected discourse across various tasks. Indeed, describing and interpreting pause patterns may also contribute to account for non-fluent aphasia, as demonstrated in previous work dealing with fluent aphasia (Butterworth, 1979). To study pause patterns in non-fluent aphasia, we performed further speech data analyses involving agrammatic speakers (N=5) and control speakers (N=9) (Sahraoui & Nespoulous, 2012). In particular, silent and filled pauses were coded and automatically computed using speech processing methods (Mac Whinney, 2000 ; Boersma & Weenink, 2015). Analyses include speech and articulatory rate, pause durations and pause distribution in relation to the elliptical style, overt errors, dysfluencies and repair strategies (corrective adaptations). In connected discourse (spontaneous and autobiographical) with more elliptical style due to more frequent omissions of grammatical morphemes, the pattern of pauses is different from narrative and descriptive discourse, as well as in sentence production. The latter type of speech output is characterized by fewer grammatical omissions with longer and more frequent pauses, what is combined with greater morpho-syntactic accuracy and complexity (that is to say less elliptical style) and with more frequent corrective adaptations. Arising from the way the speaker deals with the underlying impairment in situ, trade-offs are thus clearly made between fluency and morpho-syntactic accuracy and complexity. Even though language processing cannot perform in a safe way anymore, the agrammatic speakers rely on still operating executive functions related to pre- or post-articulatory speech monitoring in order to improve morpho-syntactic encoding.
Full Text: View/download PDF

39. Strategies for multiple feature fusion with Hierarchical HMM: Application to activity recognition from wearable audiovisual sensors

Author: Pinquier, J., Svebor Karaman, Letoupin, L., Guyot, P., Megret, R., Benois-Pineau, J., Gaestel, Y., Dartigues, J. -F, Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Image et Son, Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB)-Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Laboratoire de l'intégration, du matériau au système (IMS), Université Sciences et Technologies - Bordeaux 1-Institut Polytechnique de Bordeaux-Centre National de la Recherche Scientifique (CNRS), Institut de Santé Publique, d'Epidémiologie et de Développement (ISPED), Université Bordeaux Segalen - Bordeaux 2, Epidémiologie et Biostatistique [Bordeaux], and Université Bordeaux Segalen - Bordeaux 2-Institut de Santé Publique, d'Épidémiologie et de Développement (ISPED)-Institut National de la Santé et de la Recherche Médicale (INSERM)
Subjects: ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION, [INFO.INFO-MM]Computer Science [cs]/Multimedia [cs.MM]
Abstract: International audience; In this paper, we further develop the research on recognition of activities, in videos recorded with wearable cameras, with Hierarchical Hidden Markov Model classifiers. The visual scenes being of a strong complexity in terms of motion and visual content, good performances have been obtained using multiple visual and audio cues. The adequate fusion of features from physically different description spaces remains an open issue not only for this particular task, but in multiple problems of pattern recognition. A study of optimal fusion strategies in the HMM framework is proposed. We design and exploit early, intermediate and late fusions with emitting states in the H-HMM. The results obtained on a corpus recorded by healthy volunteers and patients in a longitudinal dementia study allow choosing optimal fusion strategies as a function of target activity.

40. Mots ou pseudo-mots ? Quel impact sur l'évaluation perceptive de l'intelligibilité ?

Author: Muriel Lalain, Alain Ghio, Corinne Fredouille, Marie Rebourg, Leslie Belleville, Margot Duval, Mathieu Balaguer, Virginie Woisard, Laboratoire Parole et Langage (LPL), Aix Marseille Université (AMU)-Centre National de la Recherche Scientifique (CNRS), Laboratoire Informatique d'Avignon (LIA), Avignon Université (AU)-Centre d'Enseignement et de Recherche en Informatique - CERI, CHU Toulouse [Toulouse], Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, INCA-C2SI Carcinologic Speech Sverity Index -n ° 2014-135, Laboratoire de phonétique de l’Université de Mons (sous l’égide de l’Institut de Recherche en Sciences et Technologies du Langage de l’UMONS), PNRIA, ANR-18-CE45-0008,RUGBI,Recherche d'unités linguistiques pertinentes pour améliorer la mesure de l'intelligibilité de la parole altérée par des troubles de production pathologique(2018), Centre Hospitalier Universitaire de Toulouse (CHU Toulouse), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), and Université de Toulouse (UT)
Subjects: Décodage Acoustico Phonétique (DAP), Production, Phonétique Clinique, [SHS.LANGUE]Humanities and Social Sciences/Linguistics, Intelligibilité, ComputingMilieux_MISCELLANEOUS, Cancer ORL
Abstract: International audience

41. Carcinologic speech severity index project: A database of speech disorder productions to assess quality of life related to speech after cancer

Author: Corine Astésano, Mathieu Balaguer, Jérôme Farinas, Corinne Fredouille, Alain Ghio, Pascal Gaillard, Laurence Giusti, Imed Laaridh, Muriel Lalain, Benoit Lepage, Julie Mauclair, Olivier Nocaudie, Julien Pinquier, Oriol Pont, Gilles Pouchoulin, Michèle Puech, Danièle Robert, Etienne Sicard, Virginie Woisard, Unité de Recherche Interdisciplinaire Octogone-Lordat (Octogone-Lordat), Université Toulouse - Jean Jaurès (UT2J), Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA), Institut de recherche en informatique de Toulouse (IRIT), Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse - Jean Jaurès (UT2J)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université Fédérale Toulouse Midi-Pyrénées-Université Toulouse 1 Capitole (UT1), Université Fédérale Toulouse Midi-Pyrénées, CHU Toulouse [Toulouse], Laboratoire Informatique d'Avignon (LIA), Avignon Université (AU)-Centre d'Enseignement et de Recherche en Informatique - CERI, Laboratoire Parole et Langage (LPL), Aix Marseille Université (AMU)-Centre National de la Recherche Scientifique (CNRS), Cognition, Langues, Langage, Ergonomie (CLLE), École pratique des hautes études (EPHE), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-Université Toulouse - Jean Jaurès (UT2J)-Centre National de la Recherche Scientifique (CNRS), Epidémiologie et analyses en santé publique : risques, maladies chroniques et handicaps (LEASP), Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées-Institut National de la Santé et de la Recherche Médicale (INSERM), Service d'ORL et Chirurgie Cervico-Facial, Hôpital de la Timone [CHU - APHM] (TIMONE), Institut National des Sciences Appliquées - Toulouse (INSA Toulouse), Institut National des Sciences Appliquées (INSA), Grantn°2014-135 from Institut National pour le CAncer (INCA) in nov. 2014, Institut National de la Santé et de la Recherche Médicale (INSERM)-Université Toulouse III - Paul Sabatier (UT3), Université Fédérale Toulouse Midi-Pyrénées-Université Fédérale Toulouse Midi-Pyrénées, Université de Toulouse (UT)-Université de Toulouse (UT), Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Institut National Polytechnique (Toulouse) (Toulouse INP), Université de Toulouse (UT)-Toulouse Mind & Brain Institut (TMBI), Université de Toulouse (UT)-Université de Toulouse (UT)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université Toulouse Capitole (UT Capitole), Université de Toulouse (UT), Centre Hospitalier Universitaire de Toulouse (CHU Toulouse), École Pratique des Hautes Études (EPHE), Université Paris sciences et lettres (PSL)-Université Paris sciences et lettres (PSL)-Université Toulouse - Jean Jaurès (UT2J), Université de Toulouse (UT)-Université de Toulouse (UT)-Centre National de la Recherche Scientifique (CNRS)-Toulouse Mind & Brain Institut (TMBI), Université de Toulouse (UT)-Université de Toulouse (UT)-Institut National de la Santé et de la Recherche Médicale (INSERM), and Institut National des Sciences Appliquées (INSA)-Université de Toulouse (UT)
Subjects: [INFO.INFO-DB]Computer Science [cs]/Databases [cs.DB], pathological speech, quality of life assessment, otorhinolaryngologic diseases, Parole, Phonétique Clinique, speech intelligibility and comprehensibility, [SHS.LANGUE]Humanities and Social Sciences/Linguistics, speech corpus, [SDV.MHEP]Life Sciences [q-bio]/Human health and pathology
Abstract: International audience; Within the framework of the Carcinologic Speech Severity Index (C2SI) InCA Project, we collected a large database of French speech recordings aiming at validating Disorder Severity Indexes. Such a database will be useful for measuring the impact of oral and pharyngeal cavity cancer on speech production. That will permit to assess patients' Quality of Life after treatment. The database is composed of audio recordings from 135 speakers and associated metadata. Several intelligibility and comprehensibility levels of speech functions have been evaluated. Acoustics and Prosody have been assessed. Perceptual evaluation rates from both naive and expert juries are being produced. Automatic analyzes are being carried out. That will provide to speech therapists objective tools to take into account the intelligibility and comprehensibility of patients which received cancer treatment (surgery and/or radiotherapy and/or chemotherapy). The aim of this paper is to justify the need of this corpus and his data collection. This corpus will be available to the scientific community through the GIS Parolotheque.

Catalog

Books, media, physical & digital resources

See catalog results

Searchworks

Select search scope, currently: Articles Catalog books, media & more in Jio Institute collections Articles journal articles & other e-resources

Search

Search Constraints

Refine your results

Search Limiters

Topic

Publication Year Range

Journal

Database

Publisher

41 results on '"Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio (IRIT-SAMoVA)"'

Search Results

Catalog

Select search scope, currently: Articles

Catalog

books, media & more in Jio Institute collections

Articles

journal articles & other e-resources