Skip to main content

Table 6 Results of ‘Visual Turing Test’ on different DCGAN- and WGAN [104]—generated brain tumor MR images presented by Han et al. [140]

From: A survey on Image Data Augmentation for Deep Learning

 

Accuracy (%)

Real selected as real

Real as synt

Synt as real

Synt as synt

T1 (DCGAN, 128 × 128)

70

26

24

6

44

Tlc (DCGAN, 128 × 128)

71

24

26

3

47

T2 (DCGAN, 128 × 128)

64

22

28

8

42

FLAIR (DCGAN, 128 × 128)

54

12

38

8

42

Concat (DCGAN, 128 × 128)

77

34

16

7

43

Concat (DCGAN, 64 × 64)

54

13

37

9

41

T1 (WGAN, 128 × 128)

64

20

30

6

44

Tlc (WGAN, 128 × 128)

55

13

37

8

42

T2 (WGAN, 128 × 128)

58

19

31

11

39

FLAIR (WGAN, 128 × 128)

62

16

34

4

46

Concat (WGAN, 128 × 128)

66

31

19

15

35

Concat (WGAN, 64 × 64)

53

18

32

15

35