Metadata on multimedia documents may help to describe their content and make their processing easier, for example by identifying events in temporal media, as well as carrying descriptive information for the overall resource. Metadata is essentially static and may be associated with, or embedded in, the multimedia contents. The aim of this paper is to present a proposal for multimedia documents annotation, based on modeling and unifying features elicited from content and structure mining. Our approach relies on the availability of annotated metadata representing segment content and structure as well as segment transcripts. Temporal and spatial operators are also taken into account when annotating documents. Any feature is identified into a descriptor called
meta-document
. These meta-documents are the basis of querying by adapted query languages.
Keywords metadata - annotation - spatiotemporal operators - querying
Ikram Amous received her Bachelor degree in business data processing of Sfax University, Tunisia, in 1998. She received her Master Degree in data processing and telecommunication from the Paul Sabatier University of Toulouse III, French in June 1999. She received her doctorate in Informatics from the Paul Sabatier University of Toulouse III in December 2003. She is currently a member at LARIM laboratory of Sfax University, Tunisia. Her research interests include semi-structured document modeling, multimedia document personalization, multimedia document annotation and querying.Anis jedidi received his Bachelor degree in business data processing from Sfax University, Tunisia, in 1999. Received his Master degree in data processing and telecommunication from Paul Sabatier University, Toulouse III, French, 2000. He is currently a Ph.D. candidate in Information System Generalized team at IRIT laboratory at Paul Sabatier University. His research interests include semi-structured document modeling, multimedia document annotation and querying.
Florence M. Sèdes is a professor of computing in the Department of Computing at Paul Sabatier University (Toulouse 3). She received a Ph.D. in Computer Science from Paul sabatier University. She has more than ten years of research experience in database and information systems. Her research interests include multimedia document, indexing and retrieving, semi-structured data and flexible querying.