An expert system for low-level image processing

Abstract

The problem approached in this thesis is how to help the novice do image processing with a rule-based system. An interactive expert system prototype, which includes the image processing functions of SADIE, has been created. A software interface between the image processing environment, SADIE, and the expert system tool, CLIPS, has been coded in standard C. Knowledge about low-level image processing (contrast enhancement and noise suppression) has been successfully implemented in the CLIPS program. The experimental results have shown that this system is efficient and friendly. Although this system was developed on a VAX under VMS, it is easily modified for most machines because it includes the portability of CLIPS. With only a minimal amount of work, we also can incorporate the system with the new SADIE C version, which has been developed in another project.