A content-based retrieval system for medical images

In this paper we describe the implementation of a medical image aid system designed for fast access to images based on their semantic content. The raw and processed images are grouped according to image type and image content and are organized in a hierarchy. The data organization in the system consists of a collection of such hierarchies that are used to efficiently store images and other related knowledge. The novelty of our system is that it accepts an arbitrary number of user-defined attributes and other features that characterize the image, as well as image processing algorithms for feature extraction. The system supports different query types and it is capable of searching images using multiple levels of symbolic and quantitative descriptions. This system provides a platform for future research on linking knowledge-based image understanding with content-based image retrieval.