Su, Linying and Sharp, Bernadette and Chibelushi, Claude (2002) Knowledge-based image understanding: A rule-based production system for X-ray segmentation. In: Proceedings of Fourth International Conference on Enterprise Information System, 3-6, April 2002, Ciudad Real - Spain .
|The full text of this publication is not available from this repository. (Contact us about this Publication)|
Image Understanding (IU) concerns the issues of how to interpret images. Knowledge-based image understanding studies the theory and techniques of computational image understanding, which uses explicit, independent knowledge about the images, such as their context and objects in them, as well as knowledge about the imaging system. Two related disciplines, Artificial Intelligence (AI) and Image Processing (IP), can contribute significantly to image understanding. A rule-based production system is a widely used knowledge representation technique, which may be used to capture various kinds of knowledge, such as perceptual, functional, and semantic knowledge, in image understanding system. This paper addresses some issues of knowledge-based approach to image understanding, presents a rule-based production system for X-ray segmentation, and proposes its expansion of incorporation with multiple knowledge sources. Here we just present the segmentation part of our research project, which aims at applying knowledge-based approach to interpret X-ray images of bone and to identify the fractured regions.
|Item Type:||Conference or workshop item (Paper)|
|Uncontrolled keywords:||Knowledge-based system, Image processing|
|Subjects:||Q Science > QA Mathematics (inc Computing science) > QA 76 Software, computer programming,|
|Divisions:||Faculties > Science Technology and Medical Studies > School of Computing > Security Group|
|Depositing User:||Mark Wheadon|
|Date Deposited:||24 Nov 2008 18:00|
|Last Modified:||13 Jul 2009 20:33|
|Resource URI:||http://kar.kent.ac.uk/id/eprint/13809 (The current URI for this page, for reference purposes)|
- Depositors only (login required):