KNOWLEDGE ACQUISITION, REPRESENTATION AND REASONING IN A GAMMA-CAMERAQUALITY-CONTROL EXPERT-SYSTEM

Gamma camera quality control (QC) requires several rest procedures whi ch involve experienced staff. In an attempt to help medical personnel and automate these tasks, a prototype expert system was developed. A l arge database of faulty QC images and associated case histories was co mpiled. These cases were used in formulating object-oriented models fo r knowledge representation and reasoning. QC image features, artefacts and hypotheses are represented as hierarchical trees, with levels cor responding to the amount of detail in the description. This paradigm a llows reasoning on various levels of abstraction. A small number of co ntrol rules derive appropriate conclusions using pattern matching tech niques. Such a modular approach overcomes the complexity and maintenan ce problems often found in traditional rule-based systems. Preliminary studies of system performance suggest that it can be used as an intel ligent QC assistant. It is hoped that the expert system can be helpful in centres lacking technical support, for example in third world coun tries.