Retina disease classification based on colour fundus images using convolutional neural networks

This paper explores Convolutional Neural Networks (CNN) as a classifier to recognize retinal images. The dataset used in this research is public STARE color image dataset comprises of 61 × 70.46 × S3, and 31×35 pixels. The dataset is categorized into 15 classes. The experimentation shows that the CNN model can achieve 80.93 percent.