Representing and Reasoning on Conceptual QueriesOver Image Databases

• Authors: Mohand-Saïd Hacid
• Venue: Journal of Intelligent Information Systems - Special issue on methodologies for intelligent information systems
• Year: 2000

The problem of content management of multimedia datatypes (e.g., image, video, graphics) is becoming increasinglyimportant with the development of advanced multimedia applications.Traditional database management systems are inadequate for thehandling of such data types. They require new techniques for queryformulation, retrieval, evaluation, and navigation. In this paper wedevelop a knowledge-based framework for modeling and retrieving imagedata by content. To represent the various aspects of an imageobject's characteristics, we propose a model which consists of threelayers: (1) Feature & Content Layer, intended to contain imagevisual features such as contours, shapes, etc.; (2) ObjectLayer, which provides the (conceptual) content dimension of images;and (3) Schema Layer, which contains the structuredabstractions of images, i.e., a general schema about the classes ofobjects represented in the object layer. We propose two abstractlanguages on the basis of description logics: one for describingknowledge of the object and schema layers, and the other, moreexpressive, for making queries. Queries can refer to the formdimension (i.e., information of the Feature & Content Layer)or to the content dimension (i.e., information of the ObjectLayer). These languages employ a variable free notation. As theamount of information contained in the previous layers may be hugeand operations performed at the Feature & Content Layer aretime-consuming, resorting to the use of materialized views to processand optimize queries may be extremely useful. For that, we propose aformal framework for testing containment of a query in a viewexpressed in our query language.