Table 1 Segmentation accuracy results compared with different models.

From: An AI method to predict pregnancy loss by extracting biological indicators from embryo ultrasound recordings in early pregnancy

Features

Model

IOU (%)

Dice (%)

Precision (%)

GS

U-Net

Attention U-net

A3F-net

89.48

92.94

97.67

98.60

96.12

98.82

94.03

94.45

98.64

YS

U-Net

Attention U-net

A3F-net

85.60

87.31

93.41

92.28

93.08

96.02

89.43

93.55

96.94

CR

U-Net

Attention U-Net

A3F-net

69.63

68.46

81.55

81.05

78.33

88.66

80.06

80.19

92.83

  1. Note: The three models are U-Net, U-Net with only attention mechanism, and A3F-net with attention fusion filtering mechanism in this paper.