Description Logics

Description logics (DL) form a family of formal knowledge representation languages, which is suitable for, among others, the expressive formalization of multimedia contents and the semantic refinement of video segmentation. The data model of the ontology, including its concept hierarchy and role hierarchy, is formalized in OWL 2 to achieve SROIQ(D) description logic expressivity while ensuring decidability. VidOnt defines ABox (Table 1), TBox (Table 2), and RBox (Table 3) axioms to support multimedia reasoning tasks (Sikos, 2016b).

Table 1. Asserting Facts with ABox Axioms

DL Syntax FOL Syntax Turtle Syntax Axiom Type
computerAnimation(Zambezia) computerAnimation(Zambezia) :Zambezia a :computerAnimation . Individual assertion
房仕龍 ≡ JackieChan 房仕龍 ≈ JackieChan :房仕龍 :sameIndividualAs :JackieChan . Individual equality

Table 2. Expressing Terminological Knowledge with TBox Axioms

DL Syntax FOL Syntax Turtle Syntax Axiom Type
liveAction ⊑ Movie ∀x(liveAction(x) → Movie(x)) :liveAction rdfs:subClassOf :Movie . Concept inclusion
remakeOf ⊑ basedOn ∀x ∀y (remakeOf(x,y) → basedOn(x,y)) :remakeOf rdfs:subPropertyOf :basedOn . Role inclusion

Table 3. Modeling Relationships between Roles with RBox Axioms

DL Syntax Turtle Syntax Axiom Type
starredIn ◦ starredIn ⊑ co-starred :co-starred owl:propertyChainAxiom (:starredIn :starredIn) . Role chain
basedOn ◦ basedOn ⊑ basedOn :basedOn a owl:TransitiveProperty . Transitive role

The integrity of the VidOnt ontology has been validated using industry-leading reasoners, such as HermiT and FaCT++.