Digital Video Libraries

etrieval by content from video dig-ital libraries requires annotation of media content at both syntactic and semantic levels. As multimedia archives become increasingly large—some broadcasting archives contain millions of hours of footage—the need for more sophisticated annotation and retrieval systems becomes in-creasingly acute. Keyword-based tagging sys-tems, such as those used by Flickr or YouTube, are simple to use but lack a common vocabulary and tag relationships, reducing their retrieval ef-fectiveness. Semantic Web technologies promise to make managing the metadata associated with these archives much easier. In the Semantic Web paradigm, ontologies become the formal tool to express concepts, their attributes, and the relationships between concepts in the do-main of interest. Indeed, ontologies can play a fundamental role in efficiently annotating content in digital video libraries because they allow association between concepts and visual data.1,2 For se-mantic annotation of visual data, there are sev-eral existing and useful ontologies, including those defined by the Dublin Core Metadata Initiative (se