Skip to main content

Table 1 Segmentation results of different methods in the in-center test scenario

From: HCA-DAN: hierarchical class-aware domain adaptive network for gastric tumor segmentation in 3D CT images

Test

Method

DSC (%) \(\uparrow\)

JI (%) \(\uparrow\)

ASD\(\downarrow\)

95HD\(\downarrow\)

D1

V-Net

50.58 ± 4.53

37.12 ± 3.65

12.94 ± 1.76

24.87 ± 5.58

3D FPN

49.46 ± 2.30

36.76 ± 2.17

13.29 ± 2.25

26.19 ± 5.57

nnU-Net

58.48 ± 4.50

44.13 ± 3.70

4.52 ± 2.18

16.01 ± 1.33

CoTr

51.53 ± 7.81

37.36 ± 6.96

12.55 ± 7.77

26.08 ± 6.42

AsTr (Ours)

59.26 ± 4.01

45.16 ± 3.56

4.84 ± 2.52

17.04 ± 1.81

D2

V-Net

47.02 ± 2.87

33.31 ± 2.47

12.72 ± 2.49

27.82 ± 2.45

3D FPN

48.60 ± 3.49

34.92 ± 2.57

14.34 ± 3.11

29.49 ± 6.62

nnU-Net

50.00 ± 6.72

35.77 ± 5.18

14.36 ± 11.89

32.43 ± 12.36

CoTr

50.52 ± 3.79

36.13 ± 3.64

8.69 ± 5.63

27.20 ± 8.33

AsTr (Ours)

55.97 ± 3.67

41.13 ± 2.52

8.24 ± 5.55

22.50 ± 6.55

D3

V-Net

39.31 ± 8.38

27.27 ± 7.73

16.61 ± 4.05

38.40 ± 8.43

3D FPN

41.92 ± 6.73

30.07 ± 5.84

20.84 ± 5.13

33.81 ± 9.21

nnU-Net

47.68 ± 4.80

34.52 ± 4.52

14.58 ± 6.80

30.40 ± 5.50

CoTr

43.13 ± 6.79

30.06 ± 5.81

14.21 ± 5.24

35.48 ± 8.69

AsTr (Ours)

48.83 ± 4.50

35.15 ± 4.10

10.52 ± 4.56

27.06 ± 3.92

D4

V-Net

59.66 ± 7.30

44.44 ± 6.52

7.35 ± 6.20

26.32 ± 11.50

3D FPN

61.51 ± 6.73

47.54 ± 6.21

7.69 ± 5.87

24.42 ± 9.42

nnU-Net

66.69 ± 6.75

51.91 ± 6.67

3.92 ± 1.84

18.10 ± 3.23

CoTr

61.26 ± 4.46

46.66 ± 3.43

7.01 ± 7.88

26.58 ± 9.99

AsTr (Ours)

67.28 ± 7.63

52.81 ± 8.41

3.19 ± 0.98

17.02 ± 6.03