FastEval Parkinsonism

FastEval Parkinsonism (https://fastevalp.cmdm.tw/) is a deep learning-driven video-based system, providing users to capture keypoints, estimate the severity, and summarize in a report.

Cite this work

@article{yang2024fasteval,
  title={FastEval Parkinsonism: an instant deep learning--assisted video-based online system for Parkinsonian motor symptom evaluation},
  author={Yang, Yu-Yuan and Ho, Ming-Yang and Tai, Chung-Hwei and Wu, Ruey-Meei and Kuo, Ming-Che and Tseng, Yufeng Jane},
  journal={NPJ Digital Medicine},
  volume={7},
  number={1},
  pages={1--13},
  year={2024},
  publisher={Nature Publishing Group}
}

Local installation for the web with all functions

Fast Installation at local by docker-compose

git clone <github-link>
# e.g. git clone [email protected]:CMDM-Lab/fasteval_parkinsonism.git
docker-compose up --build # for building

Start

docker-compose up # start server or reactivate server

Show

Open your browser and use the https://localhost:13006.
Use this link http://localhost:13006/letter_opener as your local email box so that you can receive the confirmation letter when creating an account.

Stop

docker-compose down # stop server
# or use `ctrl+C`

look at the linux environment

docker compose run web /bin/bash

Hand Predictor API

Read the document for indenpendently usage (Hand Predictor API).

Quick start with Hand Predictor API merely

Requirement: conda.

# clone the repository from github
git clone [email protected]:yuyuan871111/fast_eval_Parkinsonism.git    # by ssh
git clone https://github.com/yuyuan871111/fast_eval_Parkinsonism.git    # by html

# install and activate environment via conda
conda env create -f environment.yml # only do this for the first time
conda activate mediapipe

# change the working directory
cd ./src/lib/hand_predictor  # the path of the main script is as same as hand_predictor's README file

# find help messange
python hand_predictor.py -h

# get the testing file
wget --no-check-certificate 'https://docs.google.com/uc?export=download&id=1Jf9178R_U97Osu12cVudJvFQehtDI3U7' -O test.MOV

# main function for hand predictor API
python hand_predictor.py --wkdir_path . --seed 42 --filename test --ext MOV --hand_LR Left --hand_pos 1 --input_root_path . --output_root_path sample_output --mode single

Website development

Read the document for indenpendently usage (Website development).

Other informations

1. More information about alternative installation in TroubleShooting.
2. If you face a problem with message A server is already running. Check /home/myuser/local/app/tmp/pids/server.pid., please look at the linux environment and remove the file server.pid (shown as the codes below).

docker compose run web /bin/bash                # run the docker compose environment
rm /home/myuser/local/app/tmp/pids/server.pid   # remove the file