🧑‍⚕️
COVID-19 DETECTOR AI

Deep CNN based AI app for COVID Diagnosis using Chest X-Rays

ℹ️ This repository is a Flask version of previous iteration for the sake of deployment, which you can refer here

Live Demo : Heroku App

Tech Stack

• PyTorch (DL Model)
• OpenCV
• Flask (Backend)
• Gunicorn Server
• Heroku (PaaS)

Model Training

I have added a detailed notes on the training and inference model in this repository.

In this repo, I will only have instructions to build front end and deploy.

Instruction

To run locally,

1. Clone this repository with git clone.
2. Open a terminal inside the project directory.
3. Install dependencies using pip install -r requirements.txt
4. Run python app.py to open the app on localhost:5000.

To Deploy on Heroku

Deploying in heroku is bit of a hectic process involves a lot of bug fixing. I hope I can guide you through it. Since I used PyTorch and OpenCV in my app, heroku needs some pre-configuration.

1. Make sure you have all the requirements in requirements.py

I assume you have already created heroku account and Heroku CLI on your machine. If not, follow this tutorial.

2. Open a terminal on the project directory and create a heroku application using heroku create covid19-xray-detector
3. Since I use OpenCV Contrib Library, it requires libsm6 libxender1 and it can be installed only though apt-get. So we import buildpack in heroku, use heroku buildpacks:add --index 1 heroku-community/apt.
4. Create Aptfile, refer this
5. Now you are all set to go, deploy with a single command git push heroku master.

If you any problem with deploying your flask or django app in heroku, contact me.

I am grateful for the many community and forums online that helped me to find solutions.