MocapMe: DeepLabCut-Enhanced Neural Network for Enhanced Markerless Stability in Sit-to-Stand Motion Capture

View full-text article in PMC

. 2024 May 10;24(10):3022. doi: 10.3390/s24103022

Algorithm 1 Motion analysis using OpenPose and DeepLabCut.

1:
procedure MotionAnalysis( $v i d e o$ , $o u t p u t P a t h$ , $o p e n P o s e P a t h$ , $d e e p L a b C u t P a t h$ )
2:
$p o s e D a t a \leftarrow$ ExtractPoseData $(v i d e o, o p e n P o s e P a t h)$
3:
$d e e p L a b C u t M o d e l \leftarrow$ InitializeDeepLabCut $(d e e p L a b C u t P a t h)$
4:
for $f r a m e$ in ExtractFrames( $v i d e o, p o s e D a t a$ ) do
5:
$k e y p o i n t s \leftarrow$ ExtractKeypoints $(f r a m e, d e e p L a b C u t M o d e l)$
6:
SaveOutput( $k e y p o i n t s, o u t p u t P a t h$ )
7:
end for
8:
end procedure
9:
function ExtractPoseData( $v i d e o, o p e n P o s e P a t h$ ) ▹ Deploy OpenPose and filter output
10:
Process video with OpenPose.
11:
Parse and filter pose data based on confidence.
12:
return Filtered pose data.
13:
end function
14:
function InitializeDeepLabCut( $p a t h$ ) ▹ Set up and return DeepLabCut model
15:
Initialize and configure the DeepLabCut model.
16:
return Initialized model.
17:
end function
18:
function ExtractFrames( $v i d e o, d a t a$ ) ▹ Return frames of interest for analysis
19:
Extract frames based on pose data.
20:
return List of frames.
21:
end function
22:
function ExtractKeypoints( $f r a m e, m o d e l$ ) ▹ Apply DeepLabCut to extract keypoints
23:
Annotate frame with keypoints using the model.
24:
return Annotated keypoints.
25:
end function
26:
function SaveOutput( $d a t a, p a t h$ ) ▹ Save data to specified path
27:
Save data.
28:
end function