Multimedia information retrieval based on spatiotemporal relationships using description logics for the semantic web

Abstract

Since the latter half of the 1990s, ontology has been the main area of study in the semantic web. Ontology has been actively studied in several areas for a long time. There are many ontological applications and construction methods. In the study of the semantic web, ontologies are built using OWL capabilities. However, it is not suitable for managing spatial relationships. To represent spatial relationships, we try to create new vocabularies based on these two theories. Modeling topological relationships are accomplished using a neighborhood graph that only considers topological relatedness. In addition, we define the new vocabulary to represent the spatial relationships based on spatial description logic. In this article, we focus on multimedia information retrieval. We use the new vocabularies that were defined using these two theories for representing spatial relationships in image and video retrieval.

Utilizing Multimedia Ontologies in Video Scene Interpretation via Information Fusion and Automated Reasoning

“…Some spatial description logics implement a Region Connection Calculus, such as RCC8 (see ALC(D RCC8 ), for example [20]), while others, such as ALC(CDC), employ the Cardinal Direction Calculus (CDC) [21].…”

“…Searching for vocabularies and ontologies that contain the corresponding terms is not adequate, because the ad-hoc selection of vocabularies and ontologies will not give satisfactory results, even if the selection is limited to highquality structured data resources that have been checked for consistency. The Linked Open Vocabularies (LOV) 20 catalogue is maintained to help determine which vocabularies and ontologies to use for formal descriptions. Even though the list of rigorous criteria to meet before a vocabulary or ontology will be listed on LOV assures design quality [27], it does no guarantee that the best vocabulary will be selected for a particular scenario.…”

Conceptual Framework for Modeling Dynamic Paths from Natural Language Expressions

“…Other studies extend the research on LR relations and linguistics by exploring the use of LR relations for representing the spatial relations between the path of a moving object (represented as a line in LR relations) and a reference object (represented as a region). For example, the movement of a soccer ball towards and into the goal area (Na et al. 2006).…”

Modeling Topological Relationships between Fuzzy Spatio-Temporal Objects

“…Only when the spatial objects are embedded in the dimensions of R 2 can the topological relations be realized [22]. By eliminating those relations that are not realized, one can obtain a total of 2 relations (disjoint and equal) between two crisp points, 3 relations (disjoint, meet, contains) between one crisp point and line, 3 relations (disjoint, meet, contains) between one crisp point and region, 16 relations between two crisp lines, 13 relations between one crisp line and region [28] …”

“…And the most important model in RCC is the RCC-8, which represents 8 kinds of RCC relationships, i.e., {DC, EC, EQ, NTPP, NTPPi, TPP, TPPi, PO}, and the corresponding semantics are disjoint, meet, equal, contains, inside, covers, coveredby, and overlap, respectively. Currently, the most widely used model for analyzing topological relations is a 9-intersection model, which is based on the intersection between the parts (interior, boundary, exterior) of crisp spatial objects [28]. For two crisp spatial objects A and B, let A 0 , ∂A, A -be the interior, boundary, and exterior of A, respectively.…”