부산대학교 도서관

This paper addresses the problem of retrieving documents that contain visual information. Current visual information retrieval systems sometimes retrieve irrelevant documents or documents unrelated to the user’s query. This problem is caused by the use of low-level image descriptors; furthermore, these descriptors hardly have a semantic weight. In this work we address the image retrieval problem based on shape, since shape has a meaning by itself. On the other hand, an extension of the ontology concept, which is used in information retrieval based on text, is proposed in the image domain. Likewise, we present the Image Retrieval by Ontological Description of Shapes (IRONS) System. IRONS has been implemented and evaluated in order to analyze its utility, efficiency, and advantages comparing with well-known systems.