ABSTRACT
This chapter describes image analysis methods for computer-aided diagnosis (CAD) based on retinal fundus images. Retinal condition is described by Keith-Wagner-Barker classification or Scheie classification, which is divided into two diseases, hypertension and arteriosclerosis. The signs of low-grade diseases are related to blood vessels; thus, blood vessel segmentation techniques are very important. The blood vessel segmentation step is the most important component in retinal CAD, since it is necessary for almost all applications. The chapter introduces the artery diameter measurement as a parameter to assist diagnosis of hypertensive retinopathy. CAD techniques for diabetic retinopathy and detection of hemorrhages and microaneurysms (MA) are also described. MA is an early sign of diabetic retinopathy and has low contrast in non-contrast retinal images. The chapter then moves to explaining the techniques for detection of large cupping in the optic disc and nerve fiber layer defects for CAD of glaucoma.
