Today multimedia database systems are a most challenging area of database research. Traditional text-based information systems are increasingly extended to integrate multimedia data as for instance images, audio and video files. One special group of these extended systems are image databases for storage, management and retrieval of still images. At present, the most common technique for integrating images into a database is to store them together with some descriptive text or keywords assigned by human operators and subsequently retrieve them by matching the query texts with the stored keywords. Those texts assigned are not only very subjective and incomplete, but also very expensive. A far more promising technique -known as content-based retrieval- captures the image content automatically by visual features as colors, textures or the shape of image objects.