Content-Based Retrieval System for X-Ray Images

In this paper, we present an overview and experimental comparison of a large number of image descriptors for content-based retrieval of X-ray images. The paper concludes with recommendations which descriptors are most suitable for content-based retrieval of X-ray images. The local features invariant to scale, translation and rotation give best results. The features were integrated in content-based retrieval system for X-ray images. The system supports querying by image to find visually similar images to presented query. To employ multiple visual features to characterize image content we developed simple feature aggregation scheme.