Annotation of Pictures by Means of Graphical Ontologies
Radek Ošlejšek · International Conference on Internet Computing · 2009
This paper deals with handling semantics of a graphical data by means of ontologies and language-based interaction. The traditional concept of picture annotation is based on assigning meaning to the graphical elements. The approach introduced in this paper exploits ontologies in order to prevent chaos in terminology, to structure and filter semantics data and to automatically deduce logical consequences about the depicted objects. The presented graphical ontology framework restricts the general ontology with the aim of creating only the abstractions of the world that are valuable for the manipulation with graphical objects. The exploitation of the graphical ontology for the dialogue-based investigation of pictures is illustrated on some examples. The presented extension of the XML-based graphical format enables us to couple both the vector and raster images with the semantics defined by means of graphical ontologies.