2007-2010 / Author: Maria Sokhn
Abstract: CALIMERA is a joint collaboration PhD work with the College for Engineering and Architecture Fribourg, the European Organization Nuclear Research (CERN) and the National School of Telecommunication of Paris (TELECOM ParisTech). CALIMERA stands for Conference Advanced Level Information ManagemEnt and RetrievAl. The aim of the project is to design and implement a knowledge management framework for content based indexing and retrieval of conference video recordings. CALIMERA aims to bridge the gap between the implicit knowledge conveyed in the content resources and the explicit representation of knowledge required for efficient multimedia retrieval, access, sharing and content annotation by communities. The framework aims at providing a solution for two main tasks; the knowledge and information management of a conference video recording as well as their retrieval. We argue that the retrieval process is enhanced if the content and the context of a conference during its life cycle are taken into account for describing the resources. Therefore we propose to design a conference model based on users’ requirements in conference video recording retrieval. In order to provide a common retrieval framework for different users we have developed HELO, a High-level ModEL for cOnference composed of different scopes corresponding to a set of interpretations of the audiovisual content based on users’ retrieval needs. We studied a set of queries, put forward by different users. Based on this, we identified 8 scopes, which correspond to the most common retrieval activities of users in a scientific conference environment.