Figure 3:

Diagram of convolutional neural network (CNN) architecture. The architecture uses 7 serial convolutional 3 × 3 filters followed by the ReLU nonlinear activation function. Dropout at 50% is applied to all convolutional and fully-connected layers after the second layer. Feature maps are down sampled to 25% of the previous layer by convolutions with a stride length of two. The number of the input channels is 5. The number of activation channels in deeper layers is progressively increased from 8 to 16 to 32 to 64. Softmax is used as the activation function of the last fully connected layer.