Dynamic pictorial ontologies for video digital libraries annotation

Abstract

In this paper, we present the dynamic pictorial ontology paradigm for video annotation. Ontologies are often used to describe a given domain for different goals, including description of multimedia data. In the case of video annotation, the visual knowledge cannot be described using only abstract concepts but is more effectively represented in a visual form. To this aim, we introduce visual concepts, elicited from the data set as the most representative prototypes that specialize abstract concepts. The ontology created is intrinsically dynamic since it must embrace the perceptual and visual experience during annotation. Thus visual concepts can change, adapting to the multimedia content analyzed. Motivation for this new ontology paradigm are discussed together with a proposal of a framework for ontology creation, maintenance, and automatic annotation of video. The creation and usage of dynamic pictorial ontologies have been tested for soccer domain exploiting low level perceptual features and higher level domain features.

OVIS: ontology video surveillance indexing and retrieval system

“…They present a system that gives a solution to the semantic gap between the high-level concepts and low-level descriptors. Bertini et al [13] classify the events and the objects that are observed in video sequences by adding new instances of visual concepts to their ontology through updating mechanisms of the existing concepts. This approach used in both generic and specific domain descriptors attempts to identify visual prototypes that represent elements of visual concepts.…”

Learning ontology rules for semantic video annotation

“…A solution was presented by Bertini et al in [6], using generic and domain specific descriptors, identifying visual prototypes as representative elements of visual concepts and introducing mechanisms for their updating, as new instances of visual concepts are added to the ontology; the prototypes are used to classify events and objects observed in video sequences. Castano et al [8] have addressed the problem of temporal evolution of ontologies at the schema and visual data level.…”

“…The proposed method exploits the knowledge embedded into the ontology to learn new rules for describing video entities and events. The ontology used in this paper follows the Pictorially Enriched Ontology model [6], and includes: high-level concepts, concept properties and concept relations, used to define the semantic context of the examined domain; concept instances, with their visual descriptors, enrich the video semantic annotation. The learned rules, defined using the SWRL language, can be applied directly to an ontology defined using OWL to allow automatic semantic annotation of video sequences.…”

“…The ontology used in this work follows the Pictorially Enriched Ontology model, presented in [6]. In this model the ontology contains linguistic concepts, their relationships and instances of visual concepts.…”

Applying Semantic Web Techniques to Reservoir Engineering: Challenges and Experiences from Event Modeling

“…For instance, in bioinformatics, an event ontology has been designed to support detection of infectious disease outbreaks from the analysis of natural language text [12]. Ontologies for event definition have been defined for soccer [13] and broadcast news [14] to enhance video annotation and retrieval capabilities exploiting temporal and spatial relationships among events occurring in a video. In these two approaches Semantic Web technologies have been used for their expressiveness and for reasoning capabilities, which allow automatic specification of events categories.…”