Semantic Modeling of Digital Multimedia Babak Akhgar, Jawed Siddiqi, Fazilatur Rahman, Nazaraf Shah Faculty of ACES Sheffield Hallam University, UK Nahum Korda Straight Technology UK Raphael Attias, Norbert Benamou ORT, France Andrade Maria Teresa INESC, Portugal), Judy Dori TECHNION University, Israel Boas Hashavia IDEA, Israel Abstract - The requirement for a commonly accepted efficient mapping between multimedia metadata standards and semantic web-ontology standards is a major issue recognized by semantic multimedia research community. Though there have been several attempts to translate MPEG-7 audio descriptions to ontology languages there is very little literature that addresses issues associated with streaming video contents. In this paper we outline our plan to develop a methodology and the corresponding software implementation of mapping techniques of MPEG- 21 video items. The novelty of our effort lies in the fact that we address the complexity of video content’s metadata descriptions and its integration with the well recognized ontological standard through transparent mapping from original XML to RDF description. The validity of the proposed method and implementation detail will be verified against the MOSAICA semantic framework and its use cases. Keywords: Multimedia, Semantic Web, Ontology, MPEG, Video. 1 Introduction Content providers like television broadcasters and cultural archives are lagging behind with regard to ontology-integrated personalized systems. In order to provide personalized and context aware access to content (mostly digital multimedia contents) collected from different heterogeneous disjoint sources requires an understanding of the content as well as users using them. The Semantic Web technologies provide the means to achieve a common understanding of content and concepts needed to integrate and map content collections with enriched reasoning facilities to infer new knowledge to offer personalized service to the users. For this purpose, use of ontology-based approach has been commonly accepted by many researchers [1]. For a widespread use of ontologies in information integration and exchange, a prerequisite is the achievement of a joint standard for describing ontologies [2]. Standards activities for Semantic Web languages are mainly driven by working groups of the W3C; in particular the Semantic Web layer cake [3] proposed by Tim Berners-Lee shows the layering of the current state-of-the-art and future planned standards. While XML as a baseline allows for a syntactical description of documents, the layers RDF, Ontology and Logic are adding machine- process-able semantics - a necessary prerequisite for sharable web resources. Currently the gap in the syntactic and semantic interoperability between semantic web technologies and existing MM annotation standards recognizes that a major issue is the problem of aligning semantic web-based approaches with MPEG metadata descriptions. Choosing vocabularies to use when annotating MM is a key decision to be made as we need more than a single vocabulary to cover the image’s/object’s different relevant aspects. Many vocabularies as described by standards were developed prior to the semantic web. So the task is to translate the vocabularies to RDF or OWL. The key ISO MM standard, MPEG-7, MPEG-21 are defined using XML schema. Presently there is no commonly accepted mapping [5, 6] from XML schema to RDF or OWL (semantic web standards). This paper aims to present our plan to innovate a methodology and software developed for the interoperability of RDF with the complete MPEG-21 so that domain ontologies described in RDF can be transparently integrated with the MPEG -21 metadata. This allows applications that recognize and use the MPEG -21 constructs to make use of domain ontologies for applications like indexing, retrieval, filtering etc. resulting in more effective user retrieval and interaction with audiovisual material. The idea described in this paper will be implemented in the context of the MOSAICA architecture [22] for semantic annotation & retrieval of video items. There are several works