Content Based Multimedia Analysis Research into tools and methodologies for low- and mid-level audiovisual multimedia content analysis will be conducted. The target is to tackle the semantic gap in a bottom-up approach, i.e. adapting low-level techniques for knowledge-based multimedia analysis. For the results of these analysis algorithms suitable descriptors or description schemes for MPEG-7 or other appropriate description standards will be developed. The concrete work is structured into four activities: content structuring, moving 2D and 3D object segmentation and indexing, audio/speech processing and text analysis and content description. The information provided by this task will be used in Knowledge Extraction and Semantic Multimedia to gain higher-level information from the extracted low- to mid-level information. The developed tools will utilize interfaces and infrastructure provided by K-Space Framework Integration enabling efficient exchange of data and tools amongst different partners.   Activity 3.1 - Content Structuring Activity 3.2 - Moving 2D and 3D Object Segmentation and Indexing Activity 3.3 - Audio/Speech Processing and Text Analysis Activity 3.4 - Content Description   Partners Involved Joanneum Research Forschungsgesellschaft mbH, (JRS) Queen Mary, University of London (QMUL) Koblenz University, (Uni Ko-Ld) German Research Centre for Artificial Intelligence (DFKI) Informatics and Telematics Institute (ITI) & Centre for Research and Technology Hellas (CERTH) Dublin City University, (DCU) Centrum voor Wiskunde en Informatica, (CWI) Groupe des Ecoles des Télécommunications, (GET) Institut National de l'Audiovisuel (INA) Institut Eurecom, (EURECOM) University of Glasgow, (GU) Technische Universitat Berlin, (TUB) Ecole Polytechnique Fédérale de Lausanne, (EPFL) Contact Peter Schallauer Joanneum Research Forschungsgesellschaft mbH, (JRS)