An audio indexing and retrieval approach using a video surveillance ontology

Over the last few years, video indexing process becomes a more and more important topic in research community due to the needs of effective video data use generated by the diversity of video recorders. However, the major difficulty that arises in this research topic area is named a semantic gap problem, generated between the semantic interpretation results of low-level extracted features and the ground-truth. To resolve this problem, the ontology paradigm provides the best-proposed solution. At present, the use of these ontologies has been restricted only to video domain representation or indexing using video features and no audio ontology part has been proposed in their approach. In this paper, we consider this semantic gap problem cited above by developing a complete video surveillance ontology that regroup an audio part, used for future audio indexing purpose based essentially on SWRL (Semantic Web Rule Language) rules. Our new idea shows new prospects for future developments.