A system environment for knowledge based image analysis

H. Niemann

The purpose of image analysis is the automatic computation of a symbolic description of an image, or an image sequence. The content of the symbolic description must correspond to the requirements of a particular application and provide a description on an adequate level of abstraction. In industrial scenes, for example, a description of the quality of manufactured objects or special tasks for a robot is required. To achieve this, complex applications require the explicit representation of extensive task-specific knowledge about objects that have to be recognized, their relations to each other, and the properties of their task domain; this makes up the model. Furthermore there is a need for control algorithms to use the stored knowledge during the analysis process. The control algorithms have to be able to deal with imprecise terms of knowledge like 'dark, big, or long' and such terms must be represented by the formalism. The efficiency of an image analysis system therefore depends on both the knowledge base and the control algorithm. Besides the two modules knowledge and control, a data base for the intermediate results as well as methods of preprocessing and segmentation of images are necessary. In addition, it may be useful to have facilities for knowledge acquisition,explanation, and a transparent user interface.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: