Already in the introduction — see section 1.2 — a first clarification of the term knowledge in the context of image analysis was started. We distinguished three different aspects and five general levels of knowledge for the discussed context. Based on these distinctions and the proposed homogenous system architecture — see section 1.3 — this chapter will address the problem of knowledge representation. As a first-order approximation this requires the storage of the entities of the aspects and levels as mentioned above. Especially the representation of constraints and associations between objects and events in the real world is necessary. This is also a reminiscent of Postulate 3 in section 1.1 where the existence of structure in a complex pattern was required.
KeywordsKnowledge Representation Inference Rule Semantic Network Inference Process Segmentation Object
Unable to display preview. Download preview PDF.