🚶 🚁 🏃
Find project's website here.
Are you curious to find out how SkeletonTrackingApp works?
Watch or download the promotional video here.
SkeletonTrackingApp is a software prototype.
- Read some simple instructions on welcome screen, when you run the app.
- After
Recordingthe recorded video is saved on your desktop automatically. - After
Convert videoa text file with all joints and their respective 3D positions is saved on your desktop automatically. - Video is not required for
Text analysis. - 3D joints' positions are the
points coordinatesin meters. - Potential
NaN valuesare filled, initially, utilizinglinear interpolate [pad/forward]method and then applyingbackfillmethod. Moving averageis the default filter. Versions ofkalman filterandbutterworthcan be found here.- Default values are
depth: 1024x768,color: 1280x720 andfps=30during streaming.
- The project has been built with
python=3.7.9version onWindows 10 OS. - You must have
python3and the version to be< 3.8due to Cubemos restrictions. - You need an Intel RealSense camera
L515if you intend to use the functionalityRecording.D400serie should work as well.
To install miniconda you can follow the instructions of official website.
You need to create a new conda environment with all the essential python packages installed. You can create this environment from the provided environment.yml file in this repository in the following way:
conda env create -f environment.yml
Also, you can install requirements using the command.
pip install -r requirements.txt
It's preferable to follow the second option. In order to create a new conda environment with specific version of python you can use the following command:
conda create -n myenv python=3.7.9, wheremyenvis your desired name.
You can follow the install instructions related to Cubemos Skeleton Tracking from getting started guide or realsense installation guide.
Use the below command to run and render the app. -W ignore flag ignores useless warnings which are popped up.
python -W ignore app\run_gui.py
cd test, and then use the command pytest.
