Abstract:  

This document outlines the “Methodology for Semantics Extraction from Multimedia Content” that will be followed in the framework of the BOEMIE project. The basic aim is to describe the architecture and methods that will be used in each particular modality (text, image, video-audio) for extracting semantics from multimedia content using an ontology-driven framework. Furthermore it aims at answering the key research questions that need to be addressed in order for ontology-driven semantics extraction from multimedia content to be possible. These key issues are: (a) how ontology guides the extraction process, and (b) how the multimedia information extraction provides the means for ontology population and enrichment.