Content-based image indexing and retrieval: A syntactical approach

Abstract

This paper presents a syntactical approach to the content-based indexing and retrieval of images. Studies on the content-based indexing/retrieval of images were motivated to derive from the visual features inside images a set of objective criteria for the indexing/retrieval of the images. The existing content-based image indexing and retrieval methods, however, use in an unstructured manner the visual features such as edge elements, color histogram, texture distribution, and so on. As a result, both indexing and retrieval are not efficient and effective, compared with the text-based methods. A new hierarchical structure representation is presented in this paper. Since the hierarchical structure is generated with a data-driven approach, an objective description on an image can be obtained. The potential application of the hierarchical structure to the content-based image indexing and retrieval is discussed.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.