A5069410310- Voice and Speech Disorders
- Phonetics and Phonology Research
- Speech Recognition and Synthesis
- Natural Language Processing Techniques
- Dysphagia Assessment and Management
- Speech and Audio Processing
- French Language Learning Methods
- Neurobiology of Language and Bilingualism
- Linguistics and Discourse Analysis
- Music and Audio Processing
- Language Development and Disorders
- Diverse Musicological Studies
- Neurological disorders and treatments
- Neurogenetic and Muscular Disorders Research
- Botulinum Toxin and Related Neurological Disorders
- Stuttering Research and Treatment
- Meteorological Phenomena and Simulations
- Cerebral Palsy and Movement Disorders
- Text Readability and Simplification
- Parkinson's Disease Mechanisms and Treatments
- Assistive Technology in Communication and Mobility
Laboratoire Parole et Langage
2009-2023
Aix-Marseille Université
2010-2021
Centre National de la Recherche Scientifique
2010-2021
Laboratoire Informatique d'Avignon
2005-2019
Université d'Avignon et des Pays de Vaucluse
2008-2017
Trinity College Dublin
2012
Centre Hospitalier du Pays d'Aix
2012
Concerned with pathological voice assessment, this paper aims at characterizing dysphonia in the frequency domain for a better understanding of relating phenomena while most studies have focused only on improving classification systems diagnosis help purposes.In context, GMM-based automatic system is applied different ranges order to investigate which ones are relevant characterization.Experiment results demonstrate that low frequencies [0-3000]Hz more discrimination compared higher frequencies.
This paper addresses voice disorder assessment. It proposes an original back-and-forth methodology involving automatic classification system as well knowledge of the human experts (machine learning experts, phoneticians, and pathologists). The goal this is to bring a better understanding acoustic phenomena related dysphonia. was validated on dysphonic corpus (80 female voices), rated according GRBAS perceptual scale by expert jury. Firstly, focused frequency domain, showed interest 0–3000 Hz...
This work presents an automatic analysis of Parkinsonian speech prosody in read and spontaneous tasks.We first evaluate the reliability existing algorithms for show that these tools are efficient if a large amount data is processed.We then investigate prosodic characteristics 25 control 30 subjects (in off (non-medicated) state) both tasks characterized by longer pauses as well smaller variations pitch range, voice intensity level articulation rate; being all more marked semi-spontaneous...
This paper describes two comparative studies of voice quality assessment based on complementary approaches. The first study was undertaken 449 speakers (including 391 dysphonic patients) whose evaluated in parallel by a perceptual judgment and objective measurements acoustic aerodynamic data. Results showed that nonlinear combination 7 parameters allowed the classification 82% samples same grade as jury. second relates to adaptation Automatic Speaker Recognition (ASR) techniques pathological...
est destinée au dépôt et à la diffusion de documents scientifiques niveau recherche, publiés ou non, émanant des établissements d'enseignement recherche français étrangers, laboratoires publics privés.
Dans le cadre de la collecte des corpus oraux et multimodaux en Sciences Humaines Sociales (SHS), respect l'application du nouveau Règlement Général sur Protection Données à caractère personnel provoque un profond bouleversement dans l'organisation les méthodes travail chercheurs. C'est aussi cas Loi Jardé, relative aux recherches impliquant personne humaine, qui a étendu notion protection personnes au-delà champ recherche biomédicale. Ces deux importants changements législatif obligent...