Diabetic Retinopathy is the damage of blood vessels which can cause blindness. The detection
of diabetic retinopathy is usually carried out by ophthalmologist by analysing fundus photography.
In order to get accurate result, the fundus photography can be graded by up to four different ophthalmologists.
This standard is highly impossible to achieve in rural areas with the shortage of qualified
medical ophthalmologists. Therefore, an automated detection of diabetic retinopathy will be needed
in order to ease the burden of ophthalmologist. Furthermore, it can also be used as training module
for retinal screeners and graders. This research provides an automated detection and classification of
non-proliferative diabetic retinopathy(NPDR). More specifically, this research uses Support Vector
Machine(SVM) and Artificial Neural Network(ANN). The classification of diabetic retinopathy is
grouped into the respective stage of non-proliferative diabetic retinopathy: normal (No DR detected),
mild, and severe.