Audio-Visual Semantics: propuesta de una ontología para la descripción de secuencias audiovisuales

This paper shows the description of the conceptual aspects of audiovisual content through the Audio-Visual Semantics ontology (AVS) that allows to represent actions, characteristics and interactions between entities and / or elements with a multilevel granularity. For this, it is considered that the audiovisual pieces are composed of sequences. In these sequences it is possible to identify different happenings that can be described. Happenings are composed of a series of elements such as agents, actions, targets or other events. The use of qualifiers (specifying their scope and value) is contemplated to define qualities or attributes of the different elements that intervene in an event. The agents, actions, objects and qualifiers are not defined extensively in the ontology itself, but are referenced as concepts of SKOS vocabularies, which avoids modifying the AVS ontology and allows to define semantic relations between concepts and multilingual labeling. Currently, the AVS ontology is in development and in the validation phase through the description of audiovisual pieces and the verification of the results obtained.