william-amanabo / pnemonia-predictor Goto Github PK

Live Web App: http://child.test.woza.work/

This prototype web app uses computer vision to detect viral and bacterial pneumonia on chest x-rays. It’s a triage tool to help doctors identify high risk children fast. The app can accept batches of x-rays in jpg or png format. For each x-ray it outputs three possible conditions:

• No pneumonia
• Viral pneumonia
• Bacterial pneumonia

The model was trained on chest x-rays of children aged 1 to 5.

Model F1 score: 0.76

The process used to build and train the model is described in this Kaggle kernel:
https://www.kaggle.com/vbookshelf/child-pneumonia-diagnoser-tensorflow-js-web-app/notebook

The dataset used for the training can be found here:
https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia

All javascript, html and css files used to create the web app are available in this repo.

Needed packages :-

yarn add @babel/runtime@^7.11.2 @npmcli/move-file@^1.0.1 @tensorflow/[email protected] [email protected] [email protected] html-webpack-plugin@^4.4.1 webpack@^4.44.1 webpack-cli@^3.3.12 webpack-dev-server@^3.11.0 yarn add --dev @babel/core@^7.11.6 @babel/plugin-transform-runtime@^7.11.5

yarn add [email protected] [email protected]