Table 1 Number of participants and estimated number of labelled hours of studies using cameras to validate wearable measurements of physical activity identified in a recent systematic review36, and scoping review35.

From: Reducing annotation burden in physical activity research using vision language models

Viewpoint

No. participants

Median \(\delta t\) (s)

Hours labelled

Paper

1st

161

24

1546

19 (Oxfordshire)

1st

111

84

1078

24 (Sichuan)

1st

50

15

1218

33

1st

22

Video

11

37

3rd

22

Video

34.3

38

1st

25

20

768

34

1st

22

Video

38

39

3rd

48

Video

192

40

3rd

31

Video

31

41

  1. We recommend referring to the reviews for a more comprehensive list of validation studies. The two studies collected in Oxfordshire and Sichuan used in this work are shown at the top of this table. The estimates of the number of hours of labelled data for the timelapse studies is optimistic, since the temporal resolution of timelapse is much lower than video, resulting in periods of time that are difficult to label.