Prédiction du mouvement des lèvres à partir d'un signal de parole pour l'animation d'un avatar

Nathan Souviraà-Labastie 1
1 METISS - Speech and sound data modeling and processing
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
Abstract : Lip sync correspond to all the techniques that synchronize sounds and lips movement. It is used in several applications like virtual characters animation, playback, or dubbing. In spite of this simple and natural utilization in show business, this field remains a great challenge for scientist. The purpose of this report is to show abilities of artificial neural networks to match up audio features with lips position in real time. The aim is to provide a system more responsive than the former one based on Hidden Markov Models (HMM). It will recognize acoustic units such as phonemes or visemes. Speech recognition technics, generally off-line, use many contextual information. In our case the system will have to run on-line, we will propose solutions to overcome the lack of information caused by this constraint. Results will be compared to speech recognition state of arts and to lips animation approach already existing.
Complete list of metadatas

https://hal.inria.fr/inria-00628856
Contributor : Nathan Souviraà-Labastie <>
Submitted on : Tuesday, October 4, 2011 - 1:25:45 PM
Last modification on : Friday, November 16, 2018 - 1:22:59 AM

Identifiers

  • HAL Id : inria-00628856, version 1

Citation

Nathan Souviraà-Labastie. Prédiction du mouvement des lèvres à partir d'un signal de parole pour l'animation d'un avatar. Son [cs.SD]. 2011. ⟨inria-00628856⟩

Share

Metrics

Record views

294