A Hybrid Fuzzy Neural Networks for the Detection of Tumors in Medical Images

Abstract

In this study, we propose an approach to detect suspect zones or tumors in medical images. The idea is to define with precision the existence of different kinds of lesions using a hybrid system, which combines Fuzzy Neural Networks and Expert System. After applying a method of image segmentation to extract regions (by region growing algorithm or by mathematical morphology algorithm), the fuzzy neural networks detect the suspect regions, which are validated by an expert system to determine the nature of lesions. Some of experimental results on brain images show the feasibility of the proposed approach.