Yang Hyung Jeong, Kim Ho Young, Hur Dae Young

The Transactions of the Korea Information Processing Society (1994 ~ 2000), Vol. 5, No. 12, pp. 3025-3035, Dec. 1998
10.3745/KIPSTE.1998.5.12.3025,   PDF Download:

Abstract

In this paper, we describe the design and implementation of COIRS (COncept-Based Image Retrieval System). It differs from extant content-based image retrieval systems in that it enables users to query based on concepts - it allows users to get images conceptually relevant. A concept is basically an aggregation of primitive objects in an image. For such a concept-based image retrieval functionality, COIRS adopts an image descriptor called triple and includes a triple thesaurus used for capturing concepts. There are four facilities in COIRS: a visual image indexer, a triple thesaurus, an inverted file, and a user query interface. The visual image indexer facilitates object labeling and the specification of relative position of objects. It is an assistant tool designed to minimize manual work when indexing images. The thesaurus captures the concepts by analyzing triples, thereby extracting image semantics. The triples are then used for formulating queries as well as indexing images. The user query interface enables users to formulate queries in terms of triples or object icons. A query is evaluated by matching the triples of the query with an inverted file. The main advantages of COIRS are that 1) it is a higher level image retrieval system in comparison with other systems that retrieve images only by syntactical information such as colors, shape or texture. and 2) it provides an integrated framework into which extant content-based techonogies can be uniformly incorporated.

Statistics
Cite this article