Table 3.

Suggested setting of hyperparameters in the network

Stage	Hyperparameters	Value
Initialization	weight	1.0
	bias	0.0
	power	0.9
	max_iter	500
	learning_rate	2.5E-4
Total Loss	${\sigma }_1$ of $L_{\mathit{BCE}}$	0.6
	${\sigma }_2$ of $L_{\mathit{Dice}}$	0.4
	${\lambda }_1$ of $L_{seg4}$	1
	${\lambda }_2$ of $L_{seg3}$	0.2
	${\gamma }_1$ of $L_{adv1}$	0.005
	${\gamma }_2$ of $L_{adv2}$	0.002
	$\beta$ of $L_{\mathit{Ic}}$	0.001
Training	use_gpu	True
	is_Train	True
	batch_size	16
	max_epoch	151
	momentum	0.9
	rate_decay_weight	$(1-\frac{\mathit{iter}}{ma{x}_{i}ter})power$
	GPU RTX 2080Ti	4
Postprocessing	threshold_CCA	27