TABLE 2.

Configuration of CNN of CNN-FGR algorithm.

Layers of Network	Parameters of each layer
Convolutional layer 1 (Activation Function: ReLU)	kernel_size = 3, stride = 1 Number of feature graphs:16
Convolutional layer 2 (Activation Function: ReLU)	kernel_size = 3, stride = 2 Number of feature graphs:32
Dropout	P = 0.5
Convolutional layer 3 (Activation Function: ReLU)	kernel_size = 3, stride = 1 Number of feature graphs:32
Convolutional layer 4 (Activation Function: ReLU)	kernel_size = 3, stride = 2 Number of feature graphs: 64
Convolutional layer 5 (Activation Function: ReLU)	adaptive_avg_pool2d