Table 2 The comparison of manual and automatic wound area measurements.

From: Automatic segmentation and measurement of pressure injuries using deep learning models and a LiDAR camera

Ulcer no.

Manual

U-Net

Mask R-CNN

Area

cm2

cm2

RE (%)

cm2

RE (%)

1

1.8

2.5

40.7

2.1

19.2

2

3.9

4.1

3.6

3.9

0.8

3

16.4

14.1

13.6

14.6

11.0

4

2.7

3.2

22.3

4.7

76.6

5

14.1

9.8

30.3

16.3

16.1

6

3.9

5.5

42.4

5.7

46.5

7

43.8

37.0

15.5

36.8

16.1

8

8.8

8.7

1.4

9.4

6.5

9

8.6

9.3

8.8

9.3

8.4

10

13.4

23.0

72.3

21.0

57.1

11

17.0

26.7

56.9

27.1

59.4

12

5.1

5.5

7.1

5.4

5.8

13

8.9

11.4

27.5

13.3

48.7

14

2.7

4.6

72.9

11.3

320.3

15

6.1

7.6

25.6

15.8

159.2

*16

0.7

0.0

100.0

74.8

10,345.5

17

13.0

13.6

4.5

14.5

11.6

18

11.0

10.9

0.5

15.9

44.2

19

37.1

49.1

32.2

54.7

47.3

*20

83.2

132.0

58.7

67.0

19.5

MRE

  

31.8

 

566

SD of MRE

  

0.266

 

21.9

MRE (outliers removed)

  

26.2

 

53

SD of MRE (outliers removed)

  

0.23

 

0.75

  1. RE Relative error, MRE Mean relative error, SD Standard deviation.
  2. *The outliers.