Table 4.

Quantitative results of Different semi-supervised models trained with $Datase{t}_3$. $\mathbf{I}$ denotes the model with iteration. $\mathbf{R}$ denotes the model based on re-weighting module. $\mathbf{T}$ denotes the model based on trust module. The best results are highlighted in bold.

Method	Ground-glass opacity (GGO)				Consolidation				Average
	DSC	${\text{HD}}_{95}$	MAE	NSD	DSC	${\text{HD}}_{95}$	MAE	NSD	DSC	${\text{HD}}_{95}$	MAE	NSD
($S_1$)SSA-Net+$\mathbf{I}$	0.4058	7.3735	0.0237	0.3576	0.6251	3.1018	0.0025	0.5991	0.5155	5.2377	0.0131	0.4784
($S_2$)SSA-Net+$\mathbf{I}+\mathbf{R}$	0.4225	7.4845	0.0192	0.3605	0.6010	3.2384	0.0028	0.5661	0.5118	5.3615	0.0110	0.4633
($S_3$)SSA-Net+$\mathbf{I}+\mathbf{T}$	0.4622	5.5402	0.0117	0.4061	0.6343	3.0100	0.0026	0.5720	0.5483	4.2751	0.0072	0.4891
($S_4$)SSA-Net+$\mathbf{I}+\mathbf{T}+\mathbf{R}$	0.4654	5.9266	0.0116	0.4016	0.6562	2.9541	0.0026	0.6239	0.5608	4.4404	0.0071	0.5128