Eye disease is a severe health problem. Advanced stages of the disease may lead to vision loss. Early detection may limit the development of the severity and enhance the chance of treatment. Eye disease comes from various factors such as diabetes, increasing pressure in the eye (Glaucoma), and age-related macular degeneration. Ocular fundus 2D images are one of the most common tools used to diagnose the lining of tissue eyes. Huge data availability, increasing cases, and heavy responsibility in the health sector encourage seeking new diagnosis techniques to enhance accuracy and reduce false positive and false negative diagnoses. Computer-aided diagnosis (CAD) is the state-art-technology. This paper proposes a CAD system that combines image processing techniques and artificial intelligence. The proposed method used the green channel of fundus eye images to extract the most representative features by the trained convolutional neural network to classify five eye diseases of fundus images. The build CAD system exploits deep learning and support vector machine classifier to achieve a highly accurate model of 98% for five types of eye diseases.
