The Mila Medical research group led by Joseph Paul Cohen recently released an AI radiology assistant prototype which aims to be free tool for the entire world to identify radiographic findings and help diagnose disease in chest xray images. This tool runs in a web browser (including smartphones). It runs computation locally in the web browser so no patient data is sent on the internet and this also allows us to scale to millions of users without cost.
Note: it is not for medical use yet and is only a prototype. There are three main goals:
1. Let people play with deep learning tools
2. Show the potential of open data (needed to build a public system like this)
3. Create a tool to help teach students radiology
The system is available here: https://mlmed.org/tools/xray/ and the source code is here: https://github.com/mlmed/dl-web-xray
A technical report describing the system here: https://arxiv.org/abs/1901.11210
