Table 1 Dataset statistics.

From: Preserving privacy in surgical video analysis using a deep learning classifier to identify out-of-body scenes in endoscopic videos

Dataset

Videos

Min. duration

Max. duration

Average duration

Total frames

Out-of-body frames

(n)

(min.)

(min.)

(min.)

(n)

(n)

(%)

Training

24

35

302

112

161,870

57,203

35.34

Validation

12

26

408

132

95,157

23,388

24.58

Test

12

33

239

137

99,240

31,663

31.91

Gastric bypass

10

55

134

90

54,385

2259

4.15

Center 1

5

66

134

101

30,464

696

2.28

Center 2

5

55

117

79

23,921

1563

6.53

Cholecystectomy

20

17

88

48

58,349

5050

8.65

Center 3

5

31

88

55

16,786

2268

13.51

Center 4

5

24

88

46

13,997

885

6.32

Center 5

5

44

76

57

17,371

356

2.05

Center 6

5

17

53

34

10,195

1541

15.12