Computer assisted medical image interpretation using fuzzy logic
Norbert Gal, Vasile Stoicu-Tivadar · 2010
The information in medical imaging is structured on multiple layers: semantic and numerical. Several algorithms for shape and color detection which can be used for numerical analysis are presented in this paper. Semantic information can be extracted from numerical information using fuzzyfication. For a correct image interpretation and a diagnosis formulation several of the objects features (shape, color, orientation and spatial relations between different objects) are required. Based on these algorithms and features, this paper proposes a system that can mimic the thinking of a human medic. The system consists of a fuzzyfication engine that converts the numerical data into linguistic data. A medical knowledge database is implemented and a fuzzy inference engine is used. The medical image will be analyzed by numerical algorithms for any signs of disorder. In addition a search feature will be implemented for searching in medical databases for similar cases to confirm or infirm the diagnosis if one is discovered.