Reconfiguration layers of convolutional neural network for fundus patches classification

Convolutional neural network (CNN) is a method of supervised deep learning. The architectures including AlexNet, VGG16, VGG19, ResNet 50, ResNet101, GoogleNet, Inception-V3, Inception ResNet-V2, and Squeezenet that have 25 to 825 layers. This study aims simplify layers CNN increased accuracy for fundus patches classification. Fundus classify two categories: normal neovascularization. Data used classification MESSIDOR Retina Image Bank 2,080 patches. Results show the best 93.17% original data 99,33% augmentation using 31 It consists input layer, 7 convolutional layers, batch normalization, rectified linear unit, 6 max-pooling, fully connected softmax, output layer.