hieubkvn123/COVID-Net-demo

COVID-Net-demo

Introduction

This repository is a web demo for the COVID-Net model for COVID-19 image classification for diagnosis.

System Diagram

Set up

1. Install all requirements

pip3 install -r requirements.txt

2. Set up SQLite database

$ sqlite3 fyp.db < init.sql 

3. Heroku deployment

$ heroku git:remote -a <heroku_app_name>  # Set remote git repo to heroku app repo
$ git add . 
$ git commit -am "Commit changes"
$ git push heroku main 

4. Setting up deployment server

• 1. Go to https://github.com/hieubkvn123/COVID-Net.
• 2. Clone the repository to a public computing server.
• 3. Go to https://github.com/lindawangg/COVID-Net/blob/master/docs/models.md and download the COVIDNet-CXR-2 model and put it into the models folder of the COVID-Net repository cloned in step 1.
• 4. Run the serve_model.sh script to deploy the covidnet model.
• 5. (NOTE) If the port 8889 is not public in the computing server. Enable port 8889 using firewall-cmd:

 firewall-cmd --permanent --zone=public --add-port=8889/tcp
 firewall-cmd --reload

5. Building documentations

• 1. Requirements : Install sphinx from PYI with

pip3 install -U sphinx

• 2. Go to ./docs and build the documentation

cd docs
sphinx-apidoc -f -o . ..
make html

• 3. Open the documentation in ./docs/_build/html/index.html

TODO

• Login functionalities.

  • Create the templates and the login form.
  • Create the functionalities to retrieve data from users and compare against databse.
  • Add JWT token based authentication for login function.

• For prototype : Create records and diagnosis.

  • Create the Create Record form.
  • Create the functionalities to retrieve record data and store in database.
  • Create a function to redirect X-Ray image to computing server and retrieve result.
  • Create a function to view all records in tables.

• Preliminaries.

  • Document all of the existing functionalities.
  • Compile test cases where applicable before proceeding.

• For final presentation.

  • Finish the sorting patient records functionality,
  • Finish the search functionality.
  • Finish the view individual patient's record functionality.
  • Finish the update individual patient's record functionality.
  • Finish the delete individual patient's record functionality.

• Further functionalities.

  • Create an advanced search section to look for patient's diagnosis history.
  • Create a function to create mass-prediction for multiple profiles.

REFERENCES

• COVID-Net: A Tailored Deep Convolutional Neural Network Design for Detection of COVID-19 Cases from Chest X-Ray Images : Paper
• COVID-Net implementation : Github