Linear Prediction Residual for Efficient Diagnosis of Parkinson's Disease from Gait

This repository contains the code for all the experiments discussed in the paper.

Organization of the Repository

├── dataset
│   ├── demographics.csv
│   ├── demographics.txt
│   ├── PatientDemographics.csv
│   ├── raw
│   └── WalksDemographics.csv
├── Linear Prediction Residual for Efficient Diagnosis of Parkinson’s Disease from Gait.pdf
├── Pipeline.jpg
├── README.md
└── src
    ├── Ablation.ipynb
    ├── Comparisons
    │   ├── Baseline.ipynb
    │   ├── Batch.sh
    │   ├── Maachi et al.ipynb
    │   ├── TimeBaseline.py
    │   ├── timebaseline.txt
    │   ├── TimeMaachietal.py
    │   └── timeMaachietal.txt
    ├── DemographicsPreprocessing.ipynb
    ├── EvaluateValSplits
    │   ├── PatientLevelSplit.ipynb
    │   ├── WalkLevelSplit.ipynb
    │   └── WindowSplit.ipynb
    ├── generateLPresidual.m
    ├── Original.ipynb
    ├── timeLPresidual.sh
    ├── timeLPresidual.txt
    ├── timeOriginal.py
    ├── timeOriginal.sh
    └── timeOriginal.txt

Getting Started

• Install Matlab

• Install Python dependencies

  pip install -r requirements.txt
  

• Download dataset from Phisionet and place it in dataset/raw.

• Run the matlab script src/generateLPresidual.m to preprocess the dataset and generate LPresiduals

• View and Run Notebooks with Jupyter which can be started with the following command.

  jupyter
  

Citation

If you find this work useful please cite.

@inproceedings{alle2021linear,
  title={Linear prediction residual for efficient diagnosis of Parkinson’s disease from gait},
  author={Alle, Shanmukh and Priyakumar, U},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  pages={614--623},
  year={2021},
  organization={Springer}
}