Table 3 Evaluation Indicators ± Standard Deviation of All Competing Methods on the Chest X-Ray dataset.

From: Dual-branch attention network with deep split convolution and multi-dimensional transformers for medical image segmentation

Methods

DICE (%)

IOU (%)

RAVD (%)

ASSD (mm)

MSSD (mm)

U-Net

96.12±0.21

93.58±0.22

−1.84±0.27

2.73±0.44

11.18±0.33

U-Net++

96.63±0.13

93.24±0.21

−1.44±0.13

2.42±0.27

10.23±0.24

SwinUNet

95.32±0.16

91.98±0.20

−3.61±0.28

3.32±0.22

13.64±0.27

TransUNet

96.83±0.99

94.14±0.18

−1.33±0.22

2.16±0.26

9.32±0.22

HiFormer

96.42±0.11

93.23±0.17

−1.74±0.11

2.45±0.25

10.46±0.33

DCSAU-Net

95.19±0.24

93.34±0.37

2.26±0.86

2.19±0.35

8.48±0.29

TM-UNet

95.13±0.25

93.64±0.51

1.24±0.37

2.97±2.72

8.22±0.23

HRMedSeg

96.87±0.3

94.34±0.75

−1.47±0.4

2.26±0.91

10.16±0.56

MambaVesselNet++

96.81±0.26

94.38±0.22

−1.76±0.78

2.46±0.66

8.86±0.59

Co-Seg++

96.94±0.16

94.72±0.18

1.81±0.52

2.57±0.62

7.32±0.38

HCMFDS-Net

96.14±0.69

94.63±0.46

1.64±0.19

2.35±0.11

8.01±0.25

D3T-Net

97.78±0.08

94.84±0.14

−3.57±0.17

2.19±0.20

6.28±0.19

Back to article page