Skip to main content

Table 1 Comparative analysis of baseline models for three-vessel segmentation in full-size 3VV images

From: A deep learning framework for identifying and segmenting three vessels in fetal heart ultrasound images

Baseline

Dice

Mean

PA

Ao

SVC

IoU

HD

Dice

FCN [15]

78.81

75.84

54.55

58.61

3.95

69.73

U-Net [16]

80.75

85.10

68.40

67.56

3.65

78.08

U-Net++ [20]

82.38

86.60

69.45

69.78

3.56

79.48

Attention U-Net [21]

81.73

84.91

72.80

69.29

3.63

79.82

Deeplabv3 [22]

83.49

86.61

73.36

70.74

3.50

81.15

  1. PA: pulmonary artery; Ao: aorta; SVC: superior vena cava; Mean: the average value of the three vessels
  2. The optimal value is highlighted in bold, while the second-best value is underscored (in column)