Home page – English

Access solutions for audiovisual content offering automatic multilingual subtitling systems and experiments with a virtual French Sign Language (LSF) signer, based on the latest breakthroughs in Artificial Intelligence and Deep Learning. 


Collaborative public-private research project with 5 leading partners in their field

Project Modules

ROSETTA is based on 8 interconnected modules


From producers and broadcasters of audiovisual content to end-users, ROSETTA is targeting a wide range of stakeholders

Discover the ROSETTA project on video

ROSETTA project on video

ROSETTA project by a signing avatar

Collecting and Producing Corpora

 … to be used for machine learning approaches, film/video making, etc. 

Captioning Production Modules

In collaboration with SYSTRAN, in this module LIMSI develops neural models for the production of captioning. 

Multilingual Captioning/Subtitling Translation Module

The production of automatic captioning and subtitling into English, Spanish, Chinese, Italian and Portuguese. 

French Sign Language Generation Module – LSF

The strength of the AZee intermediate representation is to reflect the multilinear character of LSF by linking form and meaning. 

Assessment of modules

LUTIN, responsible for Module 6, will set up a classic protocol to assess the contribution of the new technologies compared to existing ones. 

In Situ and Focus Groups Tests

With the LIVING LAB approach, Module 7 aims to anticipate the success of the ROSETTA project. 

Any questions? Contact us!

Feel free to contact us for more information about the project

Subscribe to our newsletter !

Keep updated about new ROSETTA developments and achievements.