Stelios Piperidis 1,2 Infrastructure for a Multilingual Subtitle Generation System

The expansion of digital television and the increasing demand to manipulate audiovisual content underlie the need for tools and systems that will automate the multilingual subtitle generation process. In this setting the MUSA project aims at providing a system which combines speech recognition, advanced text analysis, and machine translation to help generate multilingual subtitles. In its current version the system treats English as source and target language, as far as subtitle generation is concerned, French and Greek as subtitle translation target languages. In order to train and evaluate system components, an array of application specific resources is necessary. Primary audiovisual data consist in BBC TV documentaries. For each programme the set of multifaceted multimedia parallel data captured include: the actual video, its transcript or script, English, Greek and French subtitles, and topically relevant newspaper or web-sourced extracts.