- Create a 3.9 or 3.10 conda environment environment -have tested this successfully on 3.9.16 (latest 3.9 version) and 3.10.11 (latest 3.10 version)
- Navigate to this github repo and use the command
pip install -e.
to install the dependencies from the pyproject.toml
- Open up
postprocess_GUI.py
- Scroll down to the
if __name__ == '__main__'
section at the bottom - Change the
path_to_data_folder
variable to the path to your recording -NOTE: The GUI is set up to take file paths and file names corresponding to the FreeMoCap 1.0 release. If you need to change these file paths, you can do so from theFileManager
class at the top ofpostprocess_GUI.py
- After changing the path, run the
postprocess_GUI.py
script, and the GUI will pop up
- When the GUI pops up you'll see the main menu, from which you can alter and apply various postprocessing parameters (currently mostly for interpolation/filtering)
- Press the 'Process Data and View Results' button to run the postprocessing on the skeleton, and the postprocessed skeleton should appear in the 'post-processed data' skeleton viewer
- You can move between the Filtering and Interpolation tabs at the top to more closely look at how the different processing options affect your data
- When finished, press the 'Save Out Data' button on the main menu to save the post_processed data and a .toml file of your processing parameters into the 'output_data' folder
- Default processed data file name is 'mediapipe_processed_xyz.npy' and the parameter summary name is 'postprocessing_settings.toml'