Issue No. 03 - March (2004 vol. 16)
Zbigniew Les , IEEE
<p><b>Abstract</b>—In this paper, a method of knowledge generation as part of a shape-understanding method is presented. The proposed method of knowledge generation consists of: learning the description of new a posteriori classes, learning the concept of visual objects, and generation of the visual representation of "inner” objects. The visual concept, as part of the concept of the visual object, is expressed as a set of symbolic names that refers to possible classes of shape. The visual concept can be used to find the visual similarities between different visual objects, perform visual transformations as part of visual thinking capabilities of a system, and memorize a visual object as a symbolic representation. The knowledge obtained in the process of knowledge generation is integrated with an existing knowledge of a shape understanding system and used in the explanatory process. This system of shape understanding (SUS), that is, the implementation of the shape understanding method, is designed to imitate the visual thinking capabilities of the human visual system. The SUS consists of different types of experts that perform different processing and reasoning tasks and is designed to perform visual diagnosis in medical applications.</p>
Knowledge discovery, knowledge generation, shape understanding, shape classes, visual concept.
M. Les and Z. Les, "Shape Understanding: Knowledge Generation and Learning," in IEEE Transactions on Knowledge & Data Engineering, vol. 16, no. , pp. 343-353, 2004.