Integration of Semantic Networks in Multimedia Production and Archiving
The advent of RDF (Resource Description Framework) technology in the first half of the 2000’s, followed by RDF Schema and OWL (Web Ontology Language), represented a quantum leap in the possibility of concretely using semantic networks in the most varied applications. In the field of multimedia applications, among which we can include television production and archiving in general, they are by now an indispensable ingredient for semantic enriching and contextualisation of content for a very wide range of purposes (from automatic documentation to content recommendations systems). This project aims at exploring in detail these possibilities by defining and implementing prototypical content semantic enrichment systems exploiting existing Knowledge Base and possibly extending/creating new domain ontologies.
In the RDF representation domain, and for appropriate extension in RDFS and OWL domains, knowledge is compiled as a set of assertions of type “Subject – Predicate – Object” . The terms of the assertions may refer to any entity or resource, whether it is physically existing or purely conceptual. For example, we can assert that the author of I Promessi Sposi is Alessandro Manzoni affirming that:
<dbpedia:resource/I_promessi_sposi> <dbpedia:property/autore> <dbpedia.org:resource/Alessandro_Manzoni> .
and that Alessandro manzoni is an Italian Novelist as:
<yago:Alessandro_Manzoni> <rdf:type> <yago:wikicat_19th-century_Italian_writers> .
Existing knowledge bases (eg DBPedia, Wikidata, Yago, Freebase) count millions of these “triples”, which make available and easily integrate knowledge into the most varied applications.
In the context of television archive and production, these technological components are a strategic value undoubtedly, being exploited as simple structured sources of content enrichment (documentation) as well as real-time contextual content exploration systems.
Despite the existence of the above-mentioned technologies (RDF, RDFS, OWL) and large size and deep breath semantic databases (eg DBPedia, Wikidata, Yago, Freebase), the effective integration of semantic technologies into television production and archiving is not yet a reality. The main emerging issues in this context can be mentioned: 1) coverage and up-to-dateness of semantic databases; 2) heterogeneity of facts description; 3) lack of specific domain ontologies. This project aims to analyze these aspects and to build functional prototypes that demonstrate the applicability of these technologies in real production contexts.
Related Projects
Active project
IP - based Systems and Technologies for Television Production
This programme deals with the experimentation of new IP technologies (networks, protocols, standards, equipment) in the area of television production, and particularly for studio applications.