Deep Learning Methods for Underwater Target Feature Extraction and Recognition

Table 1

CNN structure parameters.

Layer number	Type	Number of feature figures	Size of feature figures	Size of kernel	Length of step
	Input layer	1		—	—
	Convolutional layer	64			25
	Sampling layer	64			2
	Convolutional layer	128			1
	Sampling layer	128			2
	Convolutional layer	500			1
	Full connection	2000			1
	Full connection	100			1
	Full connection	3			1