ACESO

NOTE: This is made pubic for research purposes only

Dataset

Paper: A Public Data Set of Videos, Inertial Measurement Unit, and Clinical Scales of Freezing of Gait in Individuals With Parkinson's Disease During a Turning-In-Place Task Read Online

Raw data hosted on figshare

Setup

Run commands from the repository root:

source .venv/bin/activate
source .venv/bin/activate.fish # if you use fish
pip install -r requirements.txt

Running the pipeline

./run.sh runs shared preprocessing, ACESO, and baseline experiments.

1. Pre-processing

python preprocessing/parse_annotations.py    # → data/processed/annotations.json
python preprocessing/parse_labels.py         # → data/processed/labels.json
python preprocessing/parse_clinical.py       # → data/processed/clinical.json
python preprocessing/pose_yolo.py            # → cache/pose/keypoints_*.json

2. The experiment

python experiments/distill_fog/src/run_sweep.py

3. Other commands

python experiments/distill_fog/src/run_sweep.py                      # ACESO (distilled)
python experiments/distill_fog/src/run_sweep.py --variant no_distill # Baseline
python experiments/distill_fog/src/eval_merged.py --classifier alt   # → merged_w{tag}.tsv

Re-running after changes

python preprocessing/pose_yolo.py --refresh

License

See License: GPL-2.0 license