DELTIO

Objective of the project is the development of innovative techniques for the representation, analysis and extraction of semantic information for the manipulation of multimedia content, with emphasis on their application to television news bulletins. During the last few years, in parallel to unprecedented progress in computing and networking technologies, a significant growth of the volume of available multimedia content was also observed. The recent developments in digitalization, compression and content access methods provide the means for an even bigger increase of the volume of available digital content. Particularly in the area of production and distribution of news content, digital forms of representation continuously replace other traditional forms. The problem that arises is not any more the digitalization of content or production of digital content, but its organization, the ease in its access and eventually its exploitation.

To address the aforementioned problem, the broader possible automation of the analysis of multimedia content, aiming at the extraction of knowledge, is required. As an example, systems are required that can distinguish between the different parts comprising a television news bulletin (e.g. external reports, political debates, weather forecasts, etc), or even detect the thematic content (e.g. “battle scenes from Persian Gulf”). Although the techniques for the description of multimedia information have progressed significantly in the past few years, the rate of progress in the area of automatic extraction of such description from the multimedia content, particularly of semantic information, is rather slow, mainly due to the restrictions of modern algorithms for the analysis of multimedia content.

The present project will use innovative techniques of analysis that are based on a unified representation model that in combination with fuzzy inference and machine learning techniques will enable the semantic analysis of multimedia content with fusion of information available in different media (e.g. video, image, text, audio). Taking advantage of recent advances in knowledge representation developed for the Semantic Web, this project will use the OWL language for the creation and management of ontologies that will constitute the unified representation model, while on the other hand, the multimedia content will be described where possible using descriptors and description schemes of the mpeg-7 standard. For the rules, on which the fusion of meta-information from image, video and audio is based, the RuleML language is used, which defines a framework for authoring rules in the semantic web. The result the research and development activity will enable the automatic analysis of multimedia content and the extraction of knowledge, resulting in the (semi-) automatic creation of metadata, as well as the support for smart, semantic search services. The final output of the project will be a system for the analysis of television new bulletins, which will provide intelligent search functionalities in archives of digital television material.

Consortium

Athens Technology Center SA
Centre for Research and Technology Hellas CERTH-ITI
National Center of Scientific Research “Demokritos”
Institute of Computer and Communication Systems (ICCS)
Athens News Agency