Network Intrusion Detection Method Based on Multi-Scale CNN in Internet of Things

Table 1

Composition and parameters of M-CNN.

No	Network layer	Size (depth)	Step (stride)	Filling method
L1	Input layer	12 ∗ 12 ∗ 2	—	—
M2	M-module	128	S = 1	Same
C3	Conv layer	3 ∗ 3 ∗ 128	S = 1	Same
M4	M-module	256	S = 2	Same
C5	Conv layer	3 ∗ 3 ∗ 256	S = 2	Same
M6	M-module	512	S = 2	Same
C7	Conv layer	3 ∗ 3	S = 4	Same
F8	Gap layer	—	—	—
R9	Reshape layer	—	—	—
D10	LSTM layer	128	—	—
F11	Full connection layer	15	—	—