SPECIAL TRACK ON
"AI for Intelligent Access to Multimedia Data"

[ Description | ST Program | Program Committee | Topics | Submission Information ]

Description

It is nowadays widely recognized that a large disparity exists between low-level descriptors characterizing multimedia content and the richness and subjectivity of semantics in user queries and human interpretations. Audiovisual data amplify the so-called Semantic Gap.
The full exploitation of multimedia archives is only viable through methodologies, systems and technologies that allow filling this gap. The current technological situation suggests that, although multimedia information is growing in offer, its access is often tailored by requirements posed by individual technologies, providers or applications. The complexity of the management, sharing and search over multiple archives or unstructured sources (like the Web) is a burden completely left to the authors or users, no matter how rich or granular the target information is. Although technologies for indexing of AV material are now available, locating the suitable information given the scale and the heterogeneity of the current and growing archives is still a problem. As the tradition and trends in information retrieval and Web search suggests, semantic interoperability among services is the proper answer. This asks for information integration at a conceptual level and for knowledge acquisition in order to scale up to realistic sizes.

The role of AI is here twofold. On one side knowledge representation technologies are requested to support the managemenet of the overall complexity according to current standard formalisms and best practices applied in the Semantic Web. On the other hand, the typical redundancy in multimedia data opens the way to intelligent inductive technologies. Data mining and machine learning methods can be successfully applied to enrich data descriptions and increase their abstraction level. Moreover, automatic induction is the only way to scale up to the suitable dimensions. Mining the huge information and knowledge embodied by the audiovisual and partially indexed material represents an important step forward for an effective and natural interaction with distributed archives.

In this special track, researchers and practitioners from academic and industrial institutions are expected to present their work in the area of multimedia content processing and artificial intelligence. Recent theoretical and applicative results about representation, indexing, acquisition and delivery of multimedia semantics in real scenarios will be presented and discussed. The outcome will represent a possibly heterogenous but wide picture of the area that surveys current results and stimulates cross-fertilization among independent research fields.

Special Track Program

Date: 13 September, 2007, h: 9:00-13:00
Location: Sala del Terrazzo
The Final Program is now available here

Chair

Roberto Basili, University of Roma, Tor Vergata (Italy)

Program Committee

Werner Bailer, Joanneum Research (Austria)
Roberto Basili, University of Roma "Tor Vergata" (Italy) (chair)
Fabio Crestani, University of Lugano (Switzerland)
Hamish Cunningham, University of Sheffield (UK)
Daniela D'Aloisi, Fondazione Ugo Bordoni (Italy)
Alberto Messina, RAI-CRIT (Italy)
Borislav Popov, Sirma Lab. Ontotext (Bulgaria)
Raimondo Schettini University of Milano Bicocca (Italy)
Giovanni Semeraro, University of Bari (Italy)
Valentin Tablan, University of Sheffield (UK)
Daniel Teruggi, Institut National de l'Audiovisuel (France)
Fabio Vignoli, Philips Research Lab. Europe (The Netherlands)

Special Track Topics

Topics of interest include, but are not limited to:

Integration of multimedia processing and Artificial Intelligence technologies Semantic-driven multimedia indexing and retrieval Integration of content-based image/video analysis with natural language and speech processing Methods for Semantic media annotation Ontologies and Multimedia Standards bridging the multimedia and knowledge domains Cross-media machine learning and data mining Multimodal learning and reasoning techniques Management of dimensionality: low-level and high-level feature engineering Knowledge based inference over semantic media annotation Browsing large multimedia archives Scalable, semantic-driven multimedia content adaptation and summarization Interfaces and personalisation for large multimedia repositories Semantics-driven multimedia presentation generation Intelligent content, user and network-aware media engineering

Important Dates

For the important dates clik here

Submission

For on-line submission go to submission page

Instructions for Authors

The Special Track will consist of paper and poster presentations. The submissions will be managed by the Web submission portal that will be made available soon on this page. Papers must be submitted no later than 15th of April 2007.
All papers should consist of max. 12 pages (inclusive of references, tables, figures and equations), in English. PDF files should be submitted formatted according to the Springer Verlag Proceedings style. Latex style files and word document templates can be found at: LNCS Series: Author Instruction
All accepted papers and posters will be published in the AI*IA 2007 Congress proceedings on the Springer-Verlag LNAI series (previous volumes, AI*IA 2005, AI*IA 2003).