Skip to main content

Table 2 Decomposition accuracy of ML-DRSNet for different window sizes and step sizes

From: A multi-label deep residual shrinkage network for high-density surface electromyography decomposition in real-time

Window size

(data points)

Step size

(data points)

Precision

Sensitivity

F1-score

Correctly identified MU counts

20

10

0.86 ± 0.18

0.73 ± 0.29

0.76 ± 0.28

8.27 ± 9.24

20

0.88 ± 0.18

0.73 ± 0.33

0.75 ± 0.31

8.44 ± 10.15

30

0.90 ± 0.16

0.80 ± 0.28

0.81 ± 0.27

10.96 ± 10.60

40

0.89 ± 0.19

0.83 ± 0.28

0.84 ± 0.27

11.79 ± 10.48

50

0.93 ± 0.12

0.84 ± 0.25

0.85 ± 0.23

10.90 ± 10.51

60

10

0.91 ± 0.15

0.82 ± 0.24

0.84 ± 0.23

9.25 ± 8.62

20

0.89 ± 0.16

0.77 ± 0.27

0.79 ± 0.25

6.73 ± 6.84

30

0.88 ± 0.16

0.75 ± 0.28

0.77 ± 0.26

6.19 ± 6.56

40

0.85 ± 0.16

0.68 ± 0.28

0.71 ± 0.26

3.94 ± 5.41

50

0.82 ± 0.18

0.61 ± 0.30

0.65 ± 0.28

2.42 ± 3.72

100

10

0.76 ± 0.35

0.69 ± 0.40

0.70 ± 0.39

8.60 ± 8.81

20

0.74 ± 0.30

0.65 ± 0.37

0.66 ± 0.36

5.38 ± 6.54

30

0.72 ± 0.26

0.60 ± 0.33

0.61 ± 0.32

3.21 ± 3.67

40

0.71 ± 0.24

0.58 ± 0.33

0.59 ± 0.31

2.38 ± 2.89

50

0.72 ± 0.18

0.57 ± 0.28

0.58 ± 0.26

1.69 ± 2.24

140

10

0.66 ± 0.42

0.64 ± 0.46

0.64 ± 0.45

9.73 ± 9.99

20

0.64 ± 0.39

0.61 ± 0.44

0.60 ± 0.43

6.85 ± 7.85

30

0.63 ± 0.36

0.59 ± 0.42

0.58 ± 0.40

6.10 ± 6.42

40

0.65 ± 0.32

0.59 ± 0.41

0.58 ± 0.39

5.48 ± 6.00

50

0.68 ± 0.25

0.59 ± 0.37

0.58 ± 0.35

4.90 ± 5.80

  1. Values are expressed as mean ± standard deviation across 16 participants and 3 contraction intensities