Research On Key Techniques In Ontology-Based Image Retrieval Database

Currently, there are some troubles in the industry of the image production and managing multimedia resources. In this paper, image-objected material semanticized schema has been proposed aimed at the problems above. The scheme includes two aspects:one is realizing ontology repository of multimedia resource semantization, and the other is the architecture of semantic retrieval system. Furthermore, a semantic retrieval prototype system of image-objected material has been implemented by rapid prototyping, based on which the feasibility and validity of the system have been verified.This paper introduces the purpose and sense of this thesis, the current status of semantic retrieval research, the key points, the main research methods and key innovations of this research. Two-aspect theory basses:semantic web technology and multimedia retrieval. Recent semantic web architecture and its ontology hierarchy have been analyzed. Additionally, the multimedia features and shortages of content-based multimedia retrieval have been emphasized. Moreover, the shortages have been overcome by ontology technology.And then how to realize semantization problems of multimedia resource are analyzed and discussed. Image-objected multimedia resource semantization based on ontology repository is proposed. In addition, the value of multimedia resource semantization for resource retrieval has been analyzed.Finally, an image-objected material semantic retrieval system based ontology-image database is designed and implemented. This section introduces the functions, development tools and the total skeleton of this system, discusses the detailed design and implementation of main modules and analyzes the valuation method and evaluation result of this system. The conclusion of system evaluation proves the effectiveness of image-objected material semantic retrieval system.