Enhancement of automatic classification of arcus senilis-nonarcus senilis using convolutional neural networ

Cholesterol is a type of lipid found in the human body and susceptible to abnormalities. It can be detected via profiling through blood sampling. In addition, cholesterol presence "sodium ring" eye iris called corneal arcus (CA), presenting new preliminary detection method that less invasive. Therefore, this paper proposed non-invasive detecting based on convolutional neural network (CNN) model representation using 300 normal abnormal images from UBIRIS medical web images. work, contrast-limited adaptive histogram (CLAHE) unsharp masking process was applied first CA enhance quality To detect images, dataset trained tested three pre-trained CNN architectures; one created scratch, another are Resnet-50 VGG-19 architectures were fine-tuned The best result exhibited by scratch with 10-fold cross-validation produced high average accuracy at 98.81%. Thus, deeper implementation recommended future further improve localization for optometrists used their daily clinical tasks cholesterol.