Skip to main content

Table 1 Leave-one-out (LOO) cross-validation results for the best feature for intra- and interclass for the hand joint speed analysis

From: Arm movement speed assessment via a Kinect camera: A preliminary study in healthy subjects

 

Intraclass

Interclass

 

Non-Filteredf 1

Filteredf 2

LOO step

THR 1

THR 2

Error (%)

THR 1

THR 2

Error (%)

1

0.57

1.47

53.33

0.43

1.25

6.66

2

0.57

1.48

60.00

0.43

1.25

0.00

3

0.56

1.48

53.33

0.43

1.25

6.66

4

0.58

1.47

26.66

0.43

1.25

0.00

5

0.59

1.52

20.00

0.47

1.25

40.00

6

0.58

1.51

13.33

0.43

1.25

0.00

7

0.59

1.51

0.00

0.48

1.25

13.33

8

0.59

1.51

0.00

0.43

1.25

20.00

9

0.58

1.50

6.66

0.43

1.25

0.00

10

0.58

1.49

13.33

0.43

1.25

0.00

11

0.58

1.49

6.66

0.47

1.25

13.33

12

0.57

1.50

13.33

0.43

1.25

13.33

13

0.58

1.48

13.33

0.43

1.25

6.66

14

0.57

1.48

66.66

0.43

1.25

0.00

15

0.58

1.47

20.00

0.43

1.25

0.00

16

0.58

1.49

26.66

0.43

1.25

6.66

17

0.58

1.50

6.66

0.43

1.25

0.00

18

0.58

1.50

13.33

0.43

1.25

6.66

19

0.59

1.50

0.00

0.43

1.25

6.66

20

0.58

1.50

0.00

0.43

1.25

0.00

21

0.58

1.49

6.66

0.43

1.25

0.00

22

0.58

1.50

0.00

0.43

1.25

0.00

23

0.58

1.49

0.00

0.43

1.25

6.66

24

0.57

1.50

33.33

0.43

1.25

6.66

25

0.59

1.50

0.00

0.43

1.25

0.00

26

0.58

1.48

20.00

0.43

1.25

0.00

27

0.57

1.49

53.33

0.43

1.25

6.66

  1. The best feature for intra- and interclass analysis was selected based on the overall joints analysis shown in Table 1. Each LOO step means one subject is excluded from testing and the training was carried out over the remaining 26 subjects. The results of testing are displayed in the error column.