Table 1 Performances of various methods on the homologous dataset with different proportions of labels.

Labeled ratio	Methods		Evaluation metrics
Labeled ratio	Methods		DSC↑	HD₉₅ (mm)↓	JA↑
–	Baseline U-Net⁵⁵ \((\user1{\mathcal{D}}_{T} )\)		0.665	14.866	0.498
	UDA \((\user1{\mathcal{D}}_{S} ,\user1{\mathcal{D}}_{U} )\)	DOU et al.⁴⁶	0.667	9.762	0.5
	UDA \((\user1{\mathcal{D}}_{S} ,\user1{\mathcal{D}}_{U} )\)	SIFA⁵⁶	0.715	5.831	0.556
10%	MML \((\user1{\mathcal{D}}_{S} ,\user1{\mathcal{D}}_{T} )\)	DOU et al.⁴⁷	0.816	7.099	0.69
		MKD⁴⁸	0.819	6.128	0.694
		CAI et al.⁴⁹	0.798	6.312	0.664
	SSL \((\user1{\mathcal{D}}_{U} ,\user1{\mathcal{D}}_{T} )\)	FixMatch⁵⁰	0.822	3.803	0.697
		UA-MT⁵¹	0.839	5	0.722
		ConfKD⁵²	0.821	7.28	0.697
		MUE-CoT⁵³	0.832	3.606	0.713
	SSDA \((\user1{\mathcal{D}}_{S} ,\user1{\mathcal{D}}_{U} ,\user1{\mathcal{D}}_{T} )\)	Dual-Teacher + + ²⁷	0.858	5.66	0.752
		IDMNE⁵⁴	0.861	3	0.756
		SLA⁵⁷	0.875	3.162	0.777
		GFDA²⁶	0.868	3.105	0.767
		Ours	0.886	2.828	0.795
5%	MML \((\user1{\mathcal{D}}_{S} ,\user1{\mathcal{D}}_{T} )\)	DOU et al.⁴⁷	0.767	4.236	0.621
		MKD⁴⁸	0.76	6.198	0.613
		CAI et al.⁴⁹	0.785	9.899	0.647
	SSL \((\user1{\mathcal{D}}_{U} ,\user1{\mathcal{D}}_{T} )\)	FixMatch⁵⁰	0.776	3.996	0.633
		UA-MT⁵¹	0.786	5.831	0.647
		ConfKD⁵²	0.805	6.325	0.673
		MUE-CoT⁵³	0.816	4.018	0.689
	SSDA \((\user1{\mathcal{D}}_{S} ,\user1{\mathcal{D}}_{U} ,\user1{\mathcal{D}}_{T} )\)	Dual-Teacher + + ²⁷	0.82	4.243	0.696
		IDMNE⁵⁴	0.827	3.921	0.705
		SLA⁵⁷	0.828	3.317	0.707
		GFDA²⁶	0.825	4	0.702
		Ours	0.842	3.593	0.728

Quick links

Search