Ontology driven content extraction using interlingual annotation of texts in the OMNIA project
Résumé
OMNIA is an on-going project that aims to retrieve images accompanied with multilingual texts. In this paper, we pro- pose a generic method (language and do- main independent) to extract conceptual information from such texts and sponta- neous user requests. First, texts are la- belled with interlingual annotation, then a generic extractor taking a domain on- tology as a parameter extract relevant conceptual information. Implementation is also presented with a first experiment and preliminary results.