Content Based Image Retrieval Using Modified Scalable Distributed Two-Layer Data Structure

Nowadays more and more data is stored in the multimedia databases. Because describing images by keywords most often is not sufficient, performing queries in such databases is very problematic. Moreover the amount of data which has to be stored is increasing every year for example because of the rise of resolution in digital cameras. In this paper a new multimedia database structure designed for Content Based Image Retrieval is presented. It is based on Query by Shape method which decomposes an object into features like shape primitives. Since the database has to store very huge amount of data in order to provide high throughput and scalability we propose to use modified Scalable Distributed Two-layer Data Store.