DEELE-Rad: exploiting deep radiomics features in deep learning models using COVID-19 chest X-ray images

This repository contains the code used in DEELE-Rad (Deep Learning-based Radiomics) proposal.

Journal: Health Information Science and Systems

Authors: Márcus V. L. Costa, Erikson J. de Aguiar, Lucas S. Rodrigues, Caetano Traina Jr. and Agma J. M. Traina

Contents: [Paper] [Dataset] [Quickstart and Installation] [BibTeX] [Contact]

Quickstart and Installation

1. Clone the repository:

   git clone https://github.com/usmarcv/deele-rad.git
   cd deele-rad

2. The following instructions should be followed with Python 3.12 to create a Pipenv with all required installed packages. If you do not have Pipenv installed, run the following:

   pip install pipenv

   Activate the environment:

   pipenv shell

   You can install the dependencies libraries based on the Pipfile with the following command:

   pipenv sync

Running

To run our DEELE-Rad proposal, you can use two approaches:

1. Single model deep learning:

   pipenv run python main.py --model_name VGG16 --num_deep_radiomics 300 --epochs 100

2. Many models using our script.sh file:

   chmod +x script.sh
   ./script.sh

   Note: you can change the arguments to using another hyperparameters

Reference

If you use this repository, please cite the following paper:

    @inproceedings{Costa2024,
    author={Costa, Márcus V. L. and de Aguiar, Erikson J. and Rodrigues, Lucas S. and Traina, Caetano and Traina, Agma J. M.},
    journal={Health Information Science and Systems (HISS)}, 
    title={{DEELE-R}ad: exploiting deep radiomics features in deep learning models using {COVID-19} chest {X}-ray images}, 
    year={2024},
    volume={13},
    number={},
    pages={517-522},
    doi={10.1007/s13755-024-00330-6},
    url={https://link.springer.com/article/10.1007/s13755-024-00330-6}
    }

Acknowledgements

This study was financed in part by the São Paulo Research Foundation (FAPESP – grants 2016/17078-0, 2020/07200-9, 2021/08982-3, 2023/14759-0, 2023/14390-7 and 2024/09462-1), National Council for Scientific and Technological Development (CNPq – grants 152760/2021-0, 308544/2021-8, and 308738/2021-7), and Coordination for Higher Education Personnel Improvement (Finance Code 001 and grant 88887.969051/2024-00).

Contact

For more information, you can contact me by writing to marcusvlc@usp.br or LinkedIn.