/3D-pose-tracking-example

Primary LanguageJupyter Notebook

1. create a file named "data" in your working directory

2. add private data files to the "data" directory (provided by wildflower schools)

    files to request:

    - camera01_cal.json
    - camera02_cal.json
    - camera03_cal.json
    - camera04_cal.json
    - example_2d_poses_dataframe.pickle.xz
    - example_3d_pose_tracks_dataframe.pickle.xz

    once we refactor expected input to current 2d pose extraction output:
    - alphapose-results-01.json
    - alphapose-results-02.json

3. create a virtual environment running python 3.x 

    - example using an anaconda distribution:

    $ conda create -n 3d python=3.7.4
    $ conda activate 3d

    (naming the virtual env above "3d"; 
    substitute with whatever you'd like)

4. install dependencies (conda won't be able to find our custom wildflower packages)

    $ pip install -r requirements.txt

When you're finished, deactivate your virtual env

    $ conda deactivate