Classification on Digital Pathological Images of Breast Cancer Based on Deep Features of Different Levels

Table 6

CNN model structure.

Layer	Type	Num kernels	Kernel size	Stride	Activation
0	Input	3	128 × 128	—	—
1	Convolution	16	5 × 5	3	ReLU
2	Max pool	—	2 × 2	2	—
3	Convolution	32	5 × 5	1	ReLU
4	Max pool	—	2 × 2	2	—
5	Convolution	72	3 × 3	1	ReLU
6	Max pool	—	2 × 2	2	—
7	Global max pool	—	—	—	—
8	Output	2	—	—	Softmax