Table 6 Parameter performance and complexity of selected methods.

From: A semi-supervised domain adaptive medical image segmentation method based on dual-level multi-scale alignment

Methods

Parameters (M)↓

FLOPs (G)↓

DSC↑

HD95 (mm)↓

JA↑

SSL

\((\user1{\mathcal{D}}_{U} ,\user1{\mathcal{D}}_{T} )\)

FixMatch50

7.24

11.81

0.725

5.596

0.568

UA-MT51

7.24

23.62

0.735

5.383

0.582

ConfKD52

41.13

20.29

0.728

4.049

0.573

MUE-CoT53

18.70

26.36

0.762

4.849

0.615

SSDA

\((\user1{\mathcal{D}}_{S} ,\user1{\mathcal{D}}_{U} ,\user1{\mathcal{D}}_{T} )\)

Dual-Teacher +  + 27

21.72

35.43

0.783

4.289

0.644

IDMNE54

14.48

25.32

0.791

4.472

0.654

SLA57

47.36

15.46

0.807

4.098

0.677

GFDA26

14.52

35.57

0.802

4.432

0.669

Ours

14.48

24.28

0.814

4.012

0.686