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.
- 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.
- 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
- 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 hosted on Microsoft Azure Web Service
data
: folder contains the training image data for the ML modeltest_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 usespandas
to read.xlsx
file and extract information
- Clone our repository
- (Optional) Create a virtual python environment
- Run
pip install -r requirements.txt
inside the project root directory - Set up your Azure Custom Vision project and create a
credentials.json
file similar to ourcredentials_example.json
file. - Run
flask run
- The website should be up and running at
localhost:5000
- Data set from COVID-CT