A Neuro-Difference Fuzzy Technique for Automatic Segmentation of Region of Interest in Medical Imaging

Automatic selection of region of interest (ROI) becomes an important area in the fields of diagnosis and in solving storage and transmission problems. This paper introduces an automatic method for segmentation of ROI in medical imaging (MI) using neural network (NNET) and fuzzy logic. Intensity variation along radial lines, for MRI brain images and fluorescence ophthalmic images, are studied. The new proposed technique is called neuro difference fuzzy. It reduces complexity, storage space, and saves time. The detected critical points form an irregular spider hexagon filter that used to extract the ROI. Several case studies are analyzed and results are compared with other techniques