CALIMERA: Conference Advanced Level Information ManagEment & RetrievAl


 
 
The amount of recorded talks in conferences and particularly scientific ones is rapidly increasing. However, the resulting multimedia data, such as videos, lack of rich and semantic content indexing. Therefore, finding the appropriate video of a recorded talk in the increasing amount of information becomes a complex task. Based on these observations CALIMERA work consists of developing an integrated framework for content and context based video retrieval. CALIMERA proposes a novel approach that takes into account the context of a conference (logistics, conference thematic, related documents, participants, etc.) and the content of the video recording of a talk. This enables to enhance the indexing of such video, and thus its retrieval. CALIMERA is based on HELO, an integrated conference ontology that models the Conference Integrated Information.

Keywords

information retrieval, semantic web, ontology, search engine, video analysis and segmentation, information visualization, etc.

Outcomes

Several prototypes such as SMAC INVENIO INDICO & CALISEMA (see figure below). The implementation of CALIMERA framework. Demos are available on the website.

Website of the project

calimera.project.eia-fr.ch

Project Information

PhD thesis work, duration 3 years (2007-2010), in collaboration with Telecom PARISTECH, the CERN and the EIA-FR