MDHelper

Inspiration

It's an app that I envisioned for doctors and patients to use before appointments in order to save time, avoid mistakes, and to provide clarity between each other.

What it does

Their job is time-demanding and hectic. Doctors deal with lots of patients and do tons of paperwork every shift, and they even remain on-call for medical emergencies during time-offs. It's an app that allows patients to better communicate with doctors and furthermore, utilizing AI to provide extra support for the doctor when making their diagnosis.

How we built it

Built on Android with Java. The backend is running nodejs, express. Sceneform for the 3D model rendering.

Amazon services used: AWS Textract Java SDK, javascript SDK for Medical Comprehend, and AWS SES (simple email service). Blender for the 3D models.

Challenges we ran into

I wanted to use the Medical Comprehend on images for patient medication, but they didn't support images. I ended up feeding the images into AWS Textract to get the text and then used that output into comprehend and that worked perfectly. I could not get the example for Java SDK for Medical Comprehend to work (maybe because it's a bit outdated updated 3 years ago), so I ended up setting up a nodejs backend in order to use the javascript SDK and it worked well.