Adaptive Filtering of Images by Multi-agents Planning of Operators

T.K. Kalti and H. Amiri (Tunisia)


Knowledge based systems, image filtering, adaptation, multiagents system.


This paper deals with the problem of adaptation in digital image filtering. It proposes for that, an operators piloting system. The key-elements in the conception of such a system reside, in the modeling of knowledge effectuating the planning of actions to enterprise, in the selection of the operators to use for each action, and finally, in the conception of the software architecture needed to integrate all these expertises. Hence, we give a solution that proceeds, by a syntactic/semantic cooperation for the planning problem, by aptitude's scores computing for the selection and by the adoption of a multi-agents architecture for the integration of theses different kinds of knowledge. This architecture defines three conceptual levels permitting the distinction between, data coming into play, the resolution strategies manipulating them and finally tools implementing them. It uses an hybrid communication mode combining messages sending and information sharing through local dynamically-created blackboards.

Important Links:

Go Back