Table 1 Image distribution for Acriflavine and Fluorescein training and test image sets.

From: Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopy

Diagnosis categories

Acriflavine training (80%)

Acriflavine test (20%)

Fluorescein training (80%)

Fluorescein test (20%)

No dysplasia

449

108

151

35

Amalgam tattoo

25

6

0

0

Chronic inflammation

16

3

8

2

Denture associated hyperplasia

24

5

4

1

Fibroepithelial polyp

5

1

8

1

Focal papillomatosis

38

9

7

1

Hyperplasia & hyperkeratosis

325

82

119

29

Squamous papilloma

5

1

5

1

Verruciform xanthoma

10

2

0

0

Lichenoid

275

68

151

37

Oral lichenoid lesion

175

43

103

25

Oral lichen planus

100

25

48

12

Low-risk

286

69

127

31

Atypia

42

10

0

0

Low grade dysplasia

132

32

91

22

Verrucous hyperplasia

112

27

36

9

High-risk

71

17

86

22

High grade dysplasia

60

15

84

21

OSCC

11

2

2

1

Grand total

1081

262

515

125