LIMSI logo
Search 
 
    The CNRS LIMSI Directory
   
Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur
 

Spoken Language Processing Group (TLP)

RAPMAT Project


The RAPMAT (Reconnaissance Automatique de la Parole Multilingue Accentuée pour la Traduction) project aims to develop a speech-to-speech translation system targetting mobile applications in particular for under represented languages.

The main objectives of the project are to collect data representative the targetted applications (i.e. mostly conversational and accented data) and to develop acoustic and language models for this data.

As such the project is addressing the two major underlying research problems, that of speech-to-text transcription and machine translation of spoken language, as well as the development of APIs for smartphones.

Languages: English, French, Vietnamese, Urdu, Turkish, Korean, ...

The project started on June 1st 2012 and will run for two years.

Partners: LIMSI (speech-to-text, MT), Vocapia Research (speech recognition, apps)



Last modified: Saturday,02-May-15 05:43:41 CEST