A search system based on image features

We present a system platform for the content based retrieval of images. The system has resulted in a prototype with two major steps: extraction and queries. We combine manual, semi automatic and automatic approaches in order to extract image regions (colors, shapes). Regions, which are extracted semi automatically, are significant and relevant for the final user. Shapes are represented by powerful mathematical formulas, so the representation of the shape is independent of the rotation, homothecie and translation. The features extracted are modeled, using object technology, in the database management system O2. The object modeling has several advantages such as simplification of the encapsulation of data that saves developers many details of particular image hardware or software. This makes the framework portable and extensible.