Table 2 Performance of the algorithms across various cohorts in the Mobilise-D test dataset.

From: Self-supervised learning of wrist-worn daily living accelerometer data improves the automated detection of gait in older adults

 

HA

CHF

COPD

PD

PFF

OxWalk

71.75 ± 16.65

72.41 ± 8.97

72.83 ± 4.51

79.9 ± 23.82

56.29 ± 24.08

U-Net

72.57 ± 13.69

69.1 ± 4.46

56.40 ± 9.39

82.75 ± 20.31

51.53 ± 31.58

ElderNet

82.82 ± 15.08

82.21 ± 4.82

84.17 ± 7.83

83.23 ± 22.34

81.95 ± 12.08

  1. Performance is reported as the F1 score (± SD). 19 participants from the Mobilise-D used for test-set. Model performance was based on comparison with reference values (labels) obtained from the multi-sensor INDIP (INertial module with DIstance sensors and Pressure insoles) system.
  2. HA healthy adults, CHF congestive heart failure, COPD chronic obstructive pulmonary disease, PD Parkinson’s disease, PFF proximal femoral fracture.
Back to article page