An Integrated Approach to Digital Objects Storage and Retrieval

Khalil Shihab,Haider Ramadhan,Nida Al-Chalabi

doi:10.3844/jcssp.2006.683.689

Abstract

Image retrieval systems are becoming increasingly important in areas of research and commercial use. The storage of digital objects in the traditional databases is considered inadequate because of the extensive precise data required for successful retrieval. In addition, the retrieval process has been implemented using content-based image retrieval (CBIR) that relies on retrieving stored images from a collection by comparing low level features (binary form) that are automatically extracted from the images themselves. Data retrieval requires knowledge of attributes stored along with an adequate and flexible query language. For image repositories and retrieval, we noted that the integration of XML technology and case-based reasoning is more efficient and of great benefit in this area. This is mainly because users both in indexing and retrieval processes, tend to use old cases by associating images that reveal similar features. It is also because XML extends the original theory and offers a flexible approach with accurate data modelling and management tools. In this work, we also used fuzzy reasoning to convert the quantitative attributes into qualitative terms for indexing and retrieval.

Highlights

No fewer than 170,000 items had, it was universally reported, been stolen or destroyed from Baghdad’s Museum during the second Iraq war in 2003
A large number of the famous artifacts in history and treasures including the best-known ancient Mesopotamian artefacts kept in the Baghdad Museum, are believed to have been looted in the disorder following the entry of US forces into the city
The approach converts the quantitative attributes into qualitative terms for indexing and retrieval

Summary

INTRODUCTION

No fewer than 170,000 items had, it was universally reported, been stolen or destroyed from Baghdad’s Museum during the second Iraq war in 2003. As the image retrieval system must be flexible to serve variety of users with different requirements, it not possible to index images based on pre-defined categories or on simple keyword matching techniques. For this reason, free text description is normally used. If the membership function of tall, as defined, is given and if the -cut is set at 5 for tall, Image indexing and retrieval: Case attributes can be either quantitative or qualitative. Fuzzy indexing and retrieval are useful in domains where cases have quantitative attributes.

Female 8 Male

CONCLUSION