An application of NLP and audiovisual content analysis for integration of multimodal databases of current events

Abstract

The nowadays explosion of pervasive services based on the Internet makes information about current events easy and rapidly available to any user. A critical point is concerned with the possibility for the users to have a comprehensive, trustworthy and multi-source informative dossiers on their events of interest. This paper presents a completely unsupervised architecture for the automatic detection and integration of informative sources about current events based on natural language processing and audiovisual content analysis. The system provides a multimodal RSS feed integrating both audiovisual and textual contributions. © 2008 Springer-Verlag Berlin Heidelberg.