A simple pose detector which runs with the realsense framework.
Currently this is only tested on Windows and macOS. It's recommended to use Python3 (>3.7) and a virtual environment.
python install -r requirements.txtTo run an example use the basic python command to start up the script.
# start pose detection with realsense camera
python pose.pyThe landmark model currently included in MediaPipe Pose predicts the location of 33 full-body landmarks (see figure below), each with (x, y, z, visibility). Note that the z value should be discarded as the model is currently not fully trained to predict depth, but this is something we have on the roadmap.
Reference: mediapipe/solutions/pose
count- Indicates how many poses are detected (currently only0or1)- list of landmarks (
33per pose) (if pose has been detected)x- X-Position of the landmarky- Y-Position of the landmarkz- Z-Position of the landmarkvisibility- Visibility of the landmark
/mediapipe/pose [count, x, y, z, visibility, x, y, z, visibility ...]
usage: pose.py [-h] [--resolution width height] [--fps FPS]
[--model-complexity MODEL_COMPLEXITY] [--no-smooth-landmarks]
[--static-image-mode] [-mdc MIN_DETECTION_CONFIDENCE]
[-mtc MIN_TRACKING_CONFIDENCE] [--ip IP] [--port PORT]
optional arguments:
-h, --help show this help message and exit
RealSense:
--resolution width height
Resolution of the realsense stream.
--fps FPS Framerate of the realsense stream.
MediaPipe:
--model-complexity MODEL_COMPLEXITY
Set model complexity (0=Light, 1=Full, 2=Heavy).
--no-smooth-landmarks
Disable landmark smoothing.
--static-image-mode Enables static image mode.
-mdc MIN_DETECTION_CONFIDENCE, --min-detection-confidence MIN_DETECTION_CONFIDENCE
Minimum confidence value ([0.0, 1.0]) for the
detection to be considered successful.
-mtc MIN_TRACKING_CONFIDENCE, --min-tracking-confidence MIN_TRACKING_CONFIDENCE
Minimum confidence value ([0.0, 1.0]) to be considered
tracked successfully.
Network:
--ip IP The ip of the OSC server
--port PORT The port the OSC server is listening on
MIT License - Copyright (c) 2021 Florian Bruggisser
Based on the ideas of mediapipe-osc.