Table 2.

The network architecture of the ResNet generator

Layers	Input	Filter	Stride	Instance	Activation	Output
		size		normalization	function
Convolution 1	256*256*1	7*7	1	X	ReLU	256*256*64
Convolution 2	256*256*64	3*3	2	X	ReLU	128*128*128
Convolution 3	128*128*128	3*3	2	X	ReLU	64*64*256
Residual Block(RB)x7
Convolution 1 RB	64*64*256	3*3	1	X	ReLU	64*64*256
Convolution 2 RB	64*64*256	3*3	1	X	−	64*64*256
De-Convolution 1	64*64*256	3*3	2	X	ReLU	128*128*128
De-Convolution 2	128*128*128	3*3	2	X	ReLU	256*256*64
Convolution 4	256*256*64	7*7	1	−	tanh	256*256*1