Dale-Jones R & Tjahjadi T
Proc IASTED Int. Symp. A.I. Applications and Neural Networks, 1990, 73-76
This paper reports the development of a knowledge-based system for image enhancement. The system chooses an algorithm, a group of algorithms, or a combination of algorithms, suggests preprocessing which may be beneficial, and selects parameter values, where they are needed. These recommendations are made after the system has guided the user through an interactive analysis of the image, asking questions about the image content, and the form of degradation affecting it. Such questions may be answered either from direct observation by the user, or in some cases, if the information required desires analysis (e.g. the image histogram), by using image tools.
The system consists of two parts: an image processing module and an advisory program. The advisory program has a knowledge base of enhancement techniques and models of the type of degradation, represented in a combined structure of frames and rules. Object-oriented Lisp (Flavors) was chosen to implement the knowledge base.
The second part is an image processing program. This program enables the user to load images, process them using the suggested algorithms, analyse the data within the image, and to compare the effects of using different parameter values during stages of the processing.