Skip to main content

Table 4 The comprehensive comparison between our method and other research

From: Sleep postures monitoring based on capacitively coupled electrodes and deep recurrent neural networks

References

Sensor type

Number of sensors

Data size

Number of identified postures

ACC (%)

Kappa

Notes

[23]

Pressure sensors

64 × 32

13 subjects

3 (left, right and supine)

82.7

–

Large amount of sensors

[24]

Textile pressure sensors

64 × 27

12 subjects

4 (left, right, supine and prone)

97.9

0.972

Complex system design

[26]

Hard CC electrodes

13

13 subjects

4 (left, right, supine and prone)

98.4

0.967

Hard materials

[33]

Pressure sensors

14 × 32

180 subjects

3 (left, right, and s/p, i.e., supine and prone were merged as one)

94.1

0.866

Large amount of sensors

[34]

Long-narrow force sensors

16

2 subjects

3 (left, right and supine)

78.7

0.681

Data size is too small

[35]

Hard CC electrodes

20 × 15

5 subjects

3 (left, right and supine)

92.76

–

Hard materials

Proposed method

Flexible CC electrodes

3

15 subjects

3 (left, right and supine)

96.2

0.943

Non-contact soft materials