Our team members Adeeb Abbas and Stone Teele came up with this project while we were all bouncing around ideas we got while listening to the Global Management of Chronic Diseases pitch.
What it does
X-ray.ai is a software as a service that allows healthcare providers to upload chest x-rays (and ideally other diagnostic images) to a web portal and receive an accurate diagnoses of a patient's medical conditions.
How we built it
We built our server with Python and Flask. We used AWS extensively, storing and processing uploaded images in S3 and deploying our website with Elastic Beanstalk. Our machine learning was handle via Google Cloud Platform and we stored basic data (user info and authentication) with Firebase.
Challenges we ran into
Training the model was very challenging and time consuming. It was also very challenging to set up everything we needed via cloud services and make sure all the services we used with AWS communicated with each other correctly, (our team member Steve Gomberg is fantastic at DevOps and took care of all of this!).
Accomplishments that we're proud of
What we learned
We learned a lot about AWS (the good and the bad!), the extent of which we can individually function without sleep, and Flask!
What's next for X-ray.ai
There's plenty that we can still work on. So far our project is only designed to diagnose chest x-rays, but if we had more time we would allow it to take in a variety of other diagnostic images as well such as ultrasound.