This project is a WIP prototype of an experimantal skeleton-based animation classification system. Developed in Python 3.7.
It uses brute force sliding window search to compare library animations to the recorded animations and calculates the score. It then lays out library animations on the timeline to reconstruct the best match.
From the project root directory, do the following steps.
Install dependencies:
pip install requirements.txt
Set up the environment (Windows)
set PYTHONPATH=%PYTHONPATH%;%CD%/src
Set up the environment (Linux)
export PYTHONPATH=${PYTHONPATH}:${PWD}/src
Run unit tests (optional)
python -m unittest discover
Run the demo
python demos/animation_recognition_demo.py data/recorded_animation.chan data/animation_library