lma-feature-extraction

1. pip install -r requirements.txt
2. python main.py

For tests

1. pytest tests

For development

1. pip install -r requirements.txt
2. pre-commit install

Pipeline for extracting LMA features for a given bvh dataset

This schematic describes the pipeline for extracting LMA features when running:

python create_annotator.py --bvh-dir path/to/bvh-directory/ --create-annotations --downsample-step 5

The results are saved under save/<timestamp>-<bvh-directory-name>/ as csvs where each column is the component name and each row is for each frame of the motion.