Concept-based Multimedia Information Retrieval System using ontology search in Cultural heritage

The richness of Cultural Heritage and Natural History is abundant. Many of the cultural heritage collection in Library, National Archive, and Museum in the form physical object or digital format in a different type of media (text, image, audio and video). One cultural heritage object can have the relationship with other objects in different media format and do not mention query term explicitly. Using the various media format causes problems in search. A monolithic search engine like Google, Bing, Google Image, Youtube, or Findsounds only retrieve one media format. Besides, the search result of the existing search engine is less relevant and incomplete in searching cultural heritage. Several multimedia information retrieval techniques used in building the relationship using ontology like ontology based search, content-based search with ontology and hybrid search with ontology. This paper proposes Concept-based Multimedia Information Retrieval System (MIRS) with ontology using Indonesia's cultural heritage dataset to increase relevance and completeness of the system. Concept-based MIRS using manually built thesauri or by extracting latent word relationship and concept from the Ontology that provides definition and formal structure for describing the implicit and explicit concepts and its relationship in cultural heritage documentation. Ontology-based Semantic similarity measure is defined which measure the semantic relationship between document based on the likeness of their meaning. The search results indicate that the document being retrieved becomes highly relevant, more complete, enrich the keyword and in varying media formats when is compared to existing search engine results such google, bing, google image, youtube and findsounds in specific domain.