Improving Accuracy of Lung Nodule Classification Using Deep Learning with Focal Loss

Table 1

Detailed configuration of the proposed deep convolutional neural network architecture.

#	Type	Input	Kernel	Output
1	Convolutional
2	Convolutional
3	Convolutional
4	Max pooling
5	Convolutional
6	Convolutional
7	Convolutional
8	Max pooling
9	Convolutional
10	Convolutional
11	Convolutional
12	Max pooling
13	Fully connected		N/A	512
14	Dropout	512	N/A	512
15	Fully connected	512	N/A	2