Table 1.

Hyper-parameters of the GAN models used in the case study.

Hyper-parameter	Value
Activation function	ReLU (rectified linear unit)
Learning rate	0.0001
The size of minibatches	128
Epochs	150
Optimizer	Adam
Filter size	3 × 3
Channels in G	32, 32, 64, 128, 256, 256, 256, 128, 64, 32, 1
Channels in D	10, 10, 20, 40, 80, 80
Normalization	Instance normalization
Stride	2
Momentum	0.5
Attention module in D	CBAM