Table 4 The comparison of model complexity. The best results are highlighted in bold.

From: A two-stage multi-scale attention-based network for weakly supervised cataract fundus image enhancement

Method

Param (M)

Inference time (s)

CycleGAN32

57.1

11.64

CofeNet27

41.2

79.72

EnlightenGAN24

8.6

34.84

ArcNet9

54.4

59.29

PCENet12

26.6

12.63

GFENet10

89.3

25.4

TSMSA-Net(ours)

0.17

9.43

Back to article page