Detection of Diabetic Retinopathy using SVM and Random Forest Classifiers

Fundus images of the human eye play a major role in detecting retinal disorders at an early stage, in diabetic patients, so that with proper medication ophthalmologists try to reduce vision loss. These images provide ophthalmic conditions of patients and also the insights of cardiovascular defects in blood vessels of the eye. Though manual assessment is a simple process but lack of sufficient number of ophthalmologists and expertise needed is a major task to detect at an early stage. Hence with the help of machine learning algorithms through image processing techniques were gaining importance in disease detection. However, these images have to go through pre- processing techniques with filters and contrast enhancement techniques to have an easy analysis. This work tries to apply image enhancement techniques and uses machine learning algorithms to train the model and predict the accuracy of the Images. This also determines which is effectively and accurately classifying the retinal images which is having the retinal disorder caused by Diabetics known as Diabetic Retinopathy.