Automatic indexing of multimedia documents as a starting point to annotation process

Automatic text analysis widened the perspective of work on document contents by opening up the studies on the linguistic productions. In this case, we are using annotation as a case study. In our approach, annotation is defined as textual, graphic or sound information, attached to document source (text, photo, audio sequence or video sequence : multimedia). The source of our corpus is from INA databases (ie. Institut National de l’Audiovisuel, Paris). Our research task consisted of identifying what are the appropriate characteristics of a multimedia document, its context and information retrieval in the context of natural language processing (NLP), automatic indexing and knowledge representation. We discuss the crucial role of annotation process in the Knowledge Extraction tools and Management as well as in the design of Information Retrieval Systems. Our focus is more specifically on the new approach in information system design dedicated to “economic intelligence”.