Our project aims at integrating the production of speech recognition and translation into an online communication tool, which will create a true barrier-free communication platform. The project will focus on the production of French and English translations from an Arabic and English speech source. French transcripts and English translations are displayed in real time as they are decoded by speech recognition. The application also displays video of the participants in the communication.

 

The most performing current systems work by chaining the two modules of speech recognition and automatic translation. In this project, we want to extend the translation module so that it can generate multiple languages. We also want to develop an end-to-end system that avoids the problems of interfacing the modules and simplifies the in-flow processing of the speech signal.

We aim to develop a multilingual speech recognition and translation module, capable of recognising English and Arabic and translating messages into English and French, which should reduce computational resource requirements. Our platform will have the ability to be deployed on a remote server as well as embedded on a local PC.

 

 
 

Founded in 1968, SYSTRAN has developed over 50 years a unique body of knowledge and data in building language resources, parsers and translation engines. SYSTRAN has developed systems on over 140 different language pairs over the years, and continues to invest in these systems as part of the development of ever more powerful translation systems.In 2016, SYSTRAN teamed up with the Harvard University Machine Language Processing Lab to create OpenNMT, an open source tool enabling around neural translation technologies. Today, this tool is the most popular with a developer community of several thousand users, ahead of comparable tools from Google, META (Facebook), Intel, Microsoft.