Skip to main content

Table 10 Comparison of UGS-M3F with state-of-the-art methods on the FIVES dataset in terms of Number of Parameters (NP), Memory Consumption (MC), Inference Speed (IS), and GFLOPs

From: UGS-M3F: unified gated swin transformer with multi-feature fully fusion for retinal blood vessel segmentation

Method

NP

MC

IS

GFLOPs

 

(M)

(GB)

(ms)

 

U-Net [24]

31.0

12.7

18.5

65.2

GU-Net [26]

28.5

12.2

17.2

62.1

WA-Net [34]

35.2

13.7

20.1

70.8

AA-WGAN [40]

42.3

14.3

24.5

89.3

MAGF-Net [21]

48.7

14.5

22.7

94.5

NN-UNet [57]

25.8

12.1

16.8

58.4

TRANS-UNet [58]

105.0

15.7

32.5

230.5

SWIN-UNet [59]

96.4

15.4

29.3

212.4

IMFF-Net [22]

52.1

14.6

25.2

100.7

UGS-M3F

22.4

11.5

15.6

50.3