Towards an automatic translation from Arabic text to sign language

Abstract

Communication represents a very important need especially for deaf persons who use sign languages. For this reason, facilitating this communication through Information and Communication Technologies attracted many research. In this context, the Research Lab. LaTICE of the University of Tunis has been devoted during many years to develop the project Websign [1] aiming at using avatars to translate automatically a written text to sign language. The architecture of Websign is based on many modules i.e. linguistic treatment, creation sign dictionaries, animating the avatar, etc.. However, until now, Websign has not yet developed all steps of treatment related to Arabic language and in particular its linguistic step. The aim of this work is to provide an automatic translation system from an Arabic text into an animated sign language with a special focus on syntactic and semantic analysis of the input text in order to extract the maximum of inherent information that can be useful to the translation step.