Visual Domain Ontology using OWL Lite for Semantic Image Processing

In this paper, a visual domain ontology (VDO) is constructed using OWL-Lite Language. The VDO passes through two execution phases, namely, construction and inferring phases. In the construction phase, OWL classes are initialized, with reference to annotated scenes, and connected by hierarchical, spatial, and content-based relationships (presence/absence of some objects depends on other objects). In the inferring phase, the VDO is used to infer knowledge about an unknown scene. This paper aims to use a standard language, namely, OWL, to represent non-standard visual knowledge; facilitate straightforward ontology enrichment; and define the rules for inferring based on the constructed ontology. The OWL standardizes the constructed knowledge and facilitates advanced inferring because it is built on top of the first-order logic and description logic. The VDO then allows an efficient representation and reasoning of complex visual knowledge. In addition to representation, the VDO enables easy extension, sharing, and reuse of the represented visual knowledge.