Purpose

To observe a user performing knee flexion exercises and track their motion and calculate their range of motion across the entirety of the video

Getting Started

1. Install requirements in virtual environment

python -m venv venvSBP
. venvSBP/bin/activate
pip install -r requirements.txt

2. Download desired tfjs models for bodypix.

# For example, download a ResNet50-based model to ./bodypix_resnet50_float_model-stride16
# bodypix/mobilenet/float/050/model-stride8.json
$ ./get-model.sh bodypix/resnet50/float/model-stride16

3. Set path to models and image for inference in .py files

videoPath = './awesome_vid.mov'
modelPath = './bodypix_resnet50_float_model-stride16/'

4. Run the script

python evalbody_singleposemodelVIDEO.py

Acknowledgement

1. https://github.com/ajaichemmanam/simple_bodypix_python for model loading function
2. https://github.com/patlevin for support functions
3. https://github.com/likeablob for download script