Diagnosis of Diabetic Retinopathy through Retinal Fundus Images and 3D Convolutional Neural Networks with Limited Number of Samples

Table 1

Architecture hyperparameters for the proposed 3D-CNN model for binary classification task with 8 feature maps in the convolutional layer.

Layer	Filter size	Number of filters	Stride size	Dropout rate	Output size
Conv1+BN+ReLU		8	1	—
MaxPool1		—	2	—
Conv2+BN+ReLU		8	1	—
MaxPool2		—	2	—
Conv3+BN+ReLU		8	1	—
MaxPool3		—	2	—
Conv4+BN+ReLU		8	1	—
MaxPool4		—	2	—
Conv5+BN+ReLU		8	1	—
MaxPool5		—	2	—
FC 1	—	300	—	—	300
FC 2	—	150	—	—	150
FC 3	—	2	—	—	2
Dropout	—	—	—	0.1	2
Softmax	—	—	—	—	2

FC: fully connected; MaxPool: maximum pooling; BN: batch normalization; ReLU: rectified linear unit; Conv: convolutional.