Abstract
We have developed a computer-aided diagnosis system to detect abnormalities in fundus images. In Japan, ophthalmologists usually diagnose hypertensive changes by identifying narrowing arteriolar with a focus on irregularities. The purpose of this study is to develop an automated method for detecting narrowing arteriolar with a focus on irregularities in fundus images. The blood vessel candidates were detected by the density analysis method. Next, the blood vessels to be observed for diagnosis were detected by tracking the vessel candidates extending from the optic disk. In this step, the direction of a vector was determined by the angle made by a bifurcation point and a furcation. After the connectivity of the vessel segments was adjusted based on the recognized intersections, the true tree-like structure of the retinal blood vessels was established. The abnormal blood vessels were finally detected by measuring their diameters. The comparison between the results obtained using our system and the diagnostic results of ophthalmologists showed that our proposed method automatically detected an irregularity in diameter in 75.0% of all 24 narrowing arteries with a focus on irregularities in 70 fundus images. However, approximately 2.9 normal vessel segments per image were determined to be abnormal. The automated detection of narrowing arteriolar with a focus on irregularities could help ophthalmologists in diagnosing ocular diseases.
