1512 articles – 1650 references  [version française]
HAL: hal-00483851, version 1

Detailed view  Export this paper
1st International Conference on Applied Digital Human Modeling, Miami, Floride : États-Unis (2010)
A unified theoretical Bayesian model of speech communication
Clément Moulin-Frier 1, Jean-Luc Jean-Luc.Schwartz@inpg.Fr Schwartz 1, Julien Diard 2, Pierre Bessière 3
(2010-07-17)

Based on a review of models and theories in speech communication, this paper proposes an original Bayesian framework able to express each of them in a unified way. This framework allows to selectively incorporate motor processes in perception or auditory representations in production, thus implementing components of a perceptuo-motor link in speech communication processes. This provides a basis for future computational works on the joint study of perception, production and their coupling in speech communication.
1:  Grenoble Images Parole Signal Automatique (GIPSA-lab)
CNRS : UMR5216 – Université Joseph Fourier - Grenoble I – Université Pierre-Mendès-France - Grenoble II – Université Stendhal - Grenoble III – Institut Polytechnique de Grenoble - Grenoble Institute of Technology
2:  Laboratoire de psychologie et neurocognition (LPNC)
CNRS : UMR5105 – Université Pierre-Mendès-France - Grenoble II – Université Joseph Fourier - Grenoble I – Université de Savoie
3:  Laboratoire d'Informatique de Grenoble (LIG)
Université Joseph Fourier - Grenoble I – Institut Polytechnique de Grenoble - Grenoble Institute of Technology – Université Pierre-Mendès-France - Grenoble II – CNRS : UMR5217
Cognitive science/Computer science
Speech Communication – Cognitive Bayesian Modeling – Sensory-Motor interaction