Table 2 The comparative experiment results (underline: best; italics: second best; bold: lowest).

From: Cross-modal dual-domain bi-direction feature interaction network for medical imaging semantic segmentation

Method

Lung tumor DataSet

Brats2019 DataSet

Miou↑(%)

Dice↑(%)

VOE↑(%)

RVD↑(%)

Recall↑(%)

HD↓

Miou↑(%)

Dice↑(%)

VOE↑(%)

RVD↑(%)

Recall↑(%)

HD↓

Unet9

76.96

86.04

82.99

83.08

87.20

1.87

67.45

77.25

63.27

71.73

73.78

4.48

AttUnet34

78.22

87.09

84.29

84.28

87.59

1.85

65.70

74.78

58.49

71.32

72.18

4.54

SeResUnet36

77.06

85.93

81.65

82.35

86.04

1.86

66.73

77.01

65.65

73.26

74.72

4.53

UTNet38

75.14

84.53

81.21

82.61

84.99

1.94

72.84

82.83

77.92

75.86

84.09

4.33

MsTGANet35

77.98

86.78

83.48

83.79

86.92

1.86

77.55

86.25

82.05

84.34

84.35

3.94

TransUnet37

77.70

86.45

82.99

84.65

85.45

1.86

67.58

78.82

73.26

73.44

76.63

4.54

DconnNet39

77.23

86.39

83.48

82.33

88.02

1.87

78.23

86.97

85.99

85.83

88.34

4.02

Guide-Ynet3

79.72

87.98

84.60

84.84

87.63

1.78

78.23

87.19

86.55

86.88

87.09

4.04

C3TMUNet14

80.59

88.72

85.47

85.77

87.97

1.80

81.87

89.51

87.55

87.94

89.39

3.78

MEAUNet33

80.09

88.43

85.27

86.19

86.86

1.78

79.48

87.61

83.01

85.38

85.20

3.89

MdCo-Unet25

79.81

87.95

84.11

84.65

87.96

1.81

79.12

87.50

84.42

85.18

87.47

3.94

Ours

81.31

89.19

86.63

86.58

88.66

1.78

82.79

90.03

88.53

88.92

90.32

3.74

Ratiohighest(↑)

0.72%

0.47%

1.16%

0.39%

0.69%

0.00

0.92%

0.52%

0.98%

0.98%

0.93%

0.04

Ratiolowest(↑)

6.17%

4.66%

5.42%

4.23%

3.67%

0.16

17.09%

15.25%

30.04%

17.6%

18.14%

0.80

  1. Ratiohighest (↑)represents the ratio of DBW-Net experimental results to the results of the highest results among the comparison methods in Miou, Dice, VOE, RVD, and Recall. Ratiolowest (↑) represents the ratio of DBW-Net experimental results to the results of the lowest results among the comparison methods in Miou, Dice, VOE, RVD, and Recall.