Ontology for multimedia content descriptions

• Main Author: Liu, Song
• Other Authors: Chia Liang Tien
• Format: Theses and Dissertations
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Computer science and engineering::Information systems::Information storage and retrieval
• Online Access: https://hdl.handle.net/10356/40928
• Institution: Nanyang Technological University

Web ontology provides us a valuable tool for describing and representing domain knowledge on the Semantic Web. In a knowledge sharing scenario, an ontology defines the terms that are used to describe and represent an area of knowledge. Such ontology is used by people, databases and applications that need to share domain information. It is attractive to integrate Web ontology in a multimedia retrieval system, in order to facilitate fast and efficient media retrieval with an acceptable level of content interpretation. In this dissertation, as we desire to utilize ontology technology in multimedia retrieval, we have been looking at this topic from the perspective of creating high-level semantic descriptions for multimedia content, and also from the point of view of media annotation and retrieval. The novelty of our research lies in designing and integrating Web ontologies into a content-based media retrieval system. A general framework is presented for combining the ontology for still image description which provides domain knowledge descriptions, as well as an image retrieval system. A Web image retrieval system was implemented in which semantic knowledge is extracted from the Web content. A video analysis and annotation system was designed and implemented based on the content and concept descriptions that are encapsulated in the ontology for audiovisual descriptions. Moreover, we also investigated how to achieve better media content by designing and extracting new low-level and high-level feature descriptors. To achieve the content description for still images, we present an ontology for still image description. In the implemented ontology, the domain concepts, textual descriptions, and visual descriptions are encapsulated in a single ontology system. Therefore, it contains multi-modality information including both high-level and low-level descriptions for image content, together with the relevant domain knowledge.