ducngn / covid19-detector

Detect the probability of contracting COVID-19 based on Chest CT Scan Images 🦠

COVID-19 Detector

Try here

COVID-19 detector is a website that uses machine learning to give tentative predictions about an user's probability of contracting COVID-19 by looking at their chest CT-scan.

How Our Project Aids Health Care Providers

• Help doctors and healthcare workers in diagnosing patients.
• Reduce the load on testing facilities, give priority access to testing kits for people who needed it.

The Tech We Are Using

• Flask : A lightweight Python web application framework
• Jinja: for rendering templates from Flask
• Bootstrap: A front-end framework
• Docker: For containerizing and deploying our service
• A full list of dependencies in the project can be found in requirements.txt

Our machine learning model:

• Trained with Microsoft Azure Custom Vision AI
• Data set provided by COVID-CT
  • 329 Chest CT Scan Images of positive COVID-19 patients
  • 387 Chest CT Scan Images of negative COVID-19 patients

Web Host Service:

• Web hosted on Microsoft Azure Web Service

Important Project Structure:

• data: folder contains the training image data for the ML model
• test_data: folder contains the images that are separated from training data , for testing purpose.
• upload_train.py: Python script for creating, labelling, uploading, and training model on Azure.
• dataInfo.py: Script uses pandas to read .xlsx file and extract information

Hosting A Local Instance Of Our Website

1. Clone our repository
2. (Optional) Create a virtual python environment
3. Run pip install -r requirements.txt inside the project root directory
4. Set up your Azure Custom Vision project and create a credentials.json file similar to our credentials_example.json file.
5. Run flask run
6. The website should be up and running at localhost:5000

Credit:

• Data set from COVID-CT