Covidologist : A Disease Detection Web App to assist radiologists to detect the presence of COVID-19.

This project was created for KU HackFest 2022

The main purpose of this project is to automate the process of COVID-19 diagnosis and help radiologists, doctors and patients in their process by saving their precious time. So, by using a trained model on COVID chest-X-ray images, It assists medical staff and radiologists to detect the presence/sign of COVID in the chest-x-ray images reducing the workloads and speeding up the diagnosis.

Installation

Run my Project

    # clone the repo and check into the dir
    git clone https://github.com/bhimrazy/covidologist
    cd covidologist

    # Setup environment and install all the requirements
    python -m venv venv
    source venv/bin/activate
    pip install -r requirements.txt

    # Setup kaggle key or download kaggle.json key file and place it in "~/.kaggle"
    export KAGGLE_USERNAME="your kaggle username"
    export KAGGLE_KEY="your kaggle api key"

    # Download Datasets from kaggle (https://www.kaggle.com/datasets/andyczhao/covidx-cxr2)
    kaggle datasets download -d andyczhao/covidx-cxr2

    # unzip to temp folder
    unzip covidx-cxr2.zip -d temp

    # remove zip file
    rm -rf covidx-cxr2.zip


    # prepare dataset folder
    python main.py prepare

    # train model
    python main.py train

    # generate metrics
    # python main.py generate


    # Run fast api app
    cd app && uvicorn main:app --reload

Deployment Architecture

Sample images

Preview

Note: The original Heroku deployment is no longer available. You can run the project locally following the installation instructions above.

covid 19 disease detection

📚 RESOURCES

◆ PyTorch: https://pytorch.org ◆ FastAPI: https://fastapi.tiangolo.com ◆ COVIDx CXR-2 Dataset: https://www.kaggle.com/datasets/andyczhao/covidx-cxr2

Author

@bhimrazy