Automatic Detection of AMD and DME Retinal Pathologies Using Deep Learning

Table 2

Description of the created CNN and its layers.

#	Layers	Values
1	Convolution2D	Number of filters: 32; Kernel size: Activation function
2	Dropout	0.2
3	Dense	4608, activation function: ReLu
4	Dense	128, activation function: ReLu
5	Dropout	0.2
6	Dense	64, activation function: ReLu
7	GlobalAveragePooling2D	—
8	Dropout	0.5
9	Dense	3, activation function: Softmax