DIGITAL

K-Space

Knowledge Space of Semantic inference for automatic annotation and retrieval of multimedia content

K-Space ist ein nachhaltiges Netzwerk von Forschungsgruppen, das aus der Weltspitze aus dem akademischen und industriellen Umfeld besteht. Übergreifende Forschung im Bereich der semantischen Inferenz für die automatische Beschlagwortung und Suche in multimedialen Inhalten stehen dabei im Mittelpunkt. Damit soll die derzeitige Lücke zwischen automatisch generierten, low-level Beschreibungen und der Vielfältigkeit und Subjektivität in der menschlichen Interpretation von multimedialen Inhalten - semantische Lücke genannt - geschlossen werden.

Ziele
Die übergreifenden Forschungsaktivitäten in K-Space konzentrieren sich auf die folgenden drei Bereiche:

  • Semantic multimedia: Adaptierung existierender Beschreibungssprachen, Darstellung extrahierter semantischer Information, multimedialer Kontext und Reasoning und deren Integration mit den Semantic Web Technologien.
  • Content-based multimedia analysis: Werkzeuge und Methoden zur Multimedia- Signalverarbeitung, Verarbeitung von Sprache und natürlichsprachlichen Textes, multimodale Ansätze, Reduktion hochdimensionaler Datensätze, low-level Featurefusion und Datamining.
  • Knowledge extraction: Überwindung der semantischen Lücke durch gemeinsamen Nutzung der Ergebnisse der semantischen Analyse als auch der inhaltsbasierten Analyse. Aufbau einer Ontologie-Infrastruktur für multimediale Inhalte, Wissenserwerb aus multimedialen Inhalten, wissensgestützte Multimediaanalyse und Annotation, visueller Kontext und automatische Detektion von Konzepten auf hoher semantischer Ebene.

Die Forschungsgruppe Audiovisuelle Medien leitete das Arbeitspaket "Content-based multimedia analysis". Ziel dieses Arbeitspakets war es low und mid-level Algorithmen zur inhaltsbasierten Medienanalyse zu entwickeln, welche dann für die Wissenextraktion und semantische Analyse von Medien eingesetzt werden können. Speziell innerhalb dieses Arbeitspakets definierte die Forschungsgruppe entsprechende, auf MPEG-7 basierte, Beschreibungsstrukturen und notwendige Profile.

K-Space - Knowledge Space of Semantic inference for automatic annotation and retrieval of multimedia conten

The joint research activities of the network are aimed at convergence and resources optimization by exploiting important multidisciplinary aspects of multimedia knowledge extraction. This is achieved by linking research efforts over the following three research clusters underpinning the K-Space framework.

Content-based multimedia analysis: Tools and methodologies for low-level signal processing, object segmentation, audio processing, text analysis, and audiovisual content structuring and description.
Knowledge extraction: Building of a multimedia ontology infrastructure, knowledge acquisition from multimedia content, knowledge-assisted multimedia analysis, context based multimedia mining and intelligent exploitation of user relevance feedback.
Semantic multimedia: knowledge representation for multimedia, distributed semantic management of multimedia data, semantics-based interaction with multimedia and multimodal media analysis.

Main objectives of K-Space are:

  • To bring together leading European research teams to create critical mass for innovation of currently highly fragmented research groups addressing semantic inference for semiautomatic annotation and retrieval of multimedia content.
  • To build an open and expandable framework for collaborative research on knowledge acquisition based on system made up of flexible, modular and interconnected technology.
  • To disseminate the technical developments of the network across the broad research community.
  • To boost technology transfer to industry, influence and contribute to related knowledge based multimedia standardisation activities.

The DIGITAL Audiovisual Media research group lead the work package 'Content-based multimedia analysis'. Aim of this work package was to develop low and mid-level algorithms for content-based multimedia analysis which can be used for knowledge extraction and semantic multimedia analysis. A special focus was on the development of suitable descriptors and description schemes for MPEG-7 or other appropriate multimedia description standards.

DIGITAL

K-Space

Knowledge Space of Semantic inference for automatic annotation and retrieval of multimedia content

The joint research activities of the network are aimed at convergence and resources optimization by exploiting important multidisciplinary aspects of multimedia knowledge extraction. This is achieved by linking research efforts over the following three research clusters underpinning the K-Space framework.

Content-based multimedia analysis: Tools and methodologies for low-level signal processing, object segmentation, audio processing, text analysis, and audiovisual content structuring and description.
Knowledge extraction: Building of a multimedia ontology infrastructure, knowledge acquisition from multimedia content, knowledge-assisted multimedia analysis, context based multimedia mining and intelligent exploitation of user relevance feedback.
Semantic multimedia: knowledge representation for multimedia, distributed semantic management of multimedia data, semantics-based interaction with multimedia and multimodal media analysis.

Main objectives of K-Space are:

  • To bring together leading European research teams to create critical mass for innovation of currently highly fragmented research groups addressing semantic inference for semiautomatic annotation and retrieval of multimedia content.
  • To build an open and expandable framework for collaborative research on knowledge acquisition based on system made up of flexible, modular and interconnected technology.
  • To disseminate the technical developments of the network across the broad research community.
  • To boost technology transfer to industry, influence and contribute to related knowledge based multimedia standardisation activities.

The DIGITAL Audiovisual Media research group lead the work package 'Content-based multimedia analysis'. Aim of this work package was to develop low and mid-level algorithms for content-based multimedia analysis which can be used for knowledge extraction and semantic multimedia analysis. A special focus was on the development of suitable descriptors and description schemes for MPEG-7 or other appropriate multimedia description standards.