page 1  (22 pages)
2to next section

A Generalized Approach for Image Indexing and Retrieval

Based on 2-D Strings

Euripides G.M. Petrakis and Stelios C. Orphanoudakis

Department of Computer Science, University of Crete
and
Institute of Computer Science, Foundation for Research and Technology-Hellas
Heraklion, Crete, Greece

Abstract

2-D strings is one of a few representation structures originally designed for use in an IDB environment. In this paper, we propose a generalized approach for 2-D string based indexing which avoids the exhaustive search through the entire database of previous 2-D strings based techniques. The classical framework of representation of 2-D strings is also specialized to the cases of scaled and unscaled images. Then, index structures for supporting retrieval by content utilizing the 2-D string representation framework are discussed. The performance of the proposed method is evaluated using a database of simulated images and compared with the performance of existing techniques of 2-D string indexing and retrieval. The results demonstrate a very significant improvement in retrieval performance.

1 Introduction

Much attention has been given during the past few years on the design and development of Image DataBase (IDB) systems which support the archiving and the retrieval of images by content [1, 2, 3]. In such systems, images are analyzed so that descriptions of their content can be extracted and stored in the IDB together with the original images. These descriptions are then used to search the IDB and to determine which images satisfy the query selection criteria. However, the problem of retrieving images by content is difficult due to reasons related to complexity and uncertainty inherent to image analysis and interpretation tasks, the large