The Diagnosis of Diabetic Retinopathy: A Transfer Learning Approach

Diabetic Retinopathy is one of the complications of diabetes mellitus that occurs to the eye. It damages the blood vessels, which cause the leaking of the blood and other fluids due to the elevated blood glucose level. Diabetic Retinopathy is a quiet ailment that patients may not discover until abnormalities in the retina have progressed to the point that medication is difficult or impossible. It can also result in patients losing their sight completely. However, an automated screening machine may help overcome this problem by helping the ophthalmologist diagnose diabetic retinopathy patients as soon as possible. Hence, this research investigates the effectiveness of automatic screening machine by employing the Transfer Learning model such as VGG16 to extract the features and fed them to the Support Vector Machine (SVM), k-Nearest Neighbour (kNN) and Random Forest (RF) for the classification. It was shown that the VGG16-SVM pipeline displayed the most promising performance on the classification of Diabetic Retinopathy.