Table 3.

All network layers are listed with their properties.

Number	Layer Name	Layer Properties
1	Images (Input)	Size = 64 × 64 × 3
2	Conv-1	Convolutional (64 × 64 × 3 × 8) with stride 2
3	Bach Norm	Bach Normalization Operation
4	ReLU	Rectified Linear Unit
5	Max Pooling	Max-Pooling Operation (2 × 2, stride [2,2], padding = [same])
6	Dropout	50% dropout
7	Conv-2	Convolutional (32 × 32 × 3 × 16) with stride 2
8	Bach Norm	Bach Normalization Operation
9	ReLU	Rectified Linear Unit
10	Max Pooling	Max-Pooling Operation (2 × 2, stride [2,2])
11	Dropout	50% dropout
12	Conv-3	Convolutional (16 × 16 × 3 × 32, stride 2, padding = [0,0,0,0])
13	Bach Norm	Bach Normalization Operation
14	ReLU	Rectified Linear Unit
15	Max Pooling	Max-Pooling Operation (2 × 2, stride [2,2], padding = [same])
16	Dropout	50%
17	Conv-4	Convolutional (8 × 8 × 3 × 64) with stride 2
18	Bach Norm	Bach Normalization Operation
19	ReLU	Rectified Linear Unit
20	Max Pooling	Max-Pooling Operation (2 × 2, stride [2,2], padding = [same])
21	Dropout	50%
22	Fully Connected	512 hidden neurons in first hidden layer and 1024 in second hidden layer
23	Functions	tanh on first and second hidden layers neurons, and sigmoid on the output layer neuron.
24	Classification	Output (Normal or abnormal)
25	Loss	Binary Cross-entropy