My MedicAI

Inspiration 💡

The best years of my life so far were spent working in a Level II trauma center. I worked along some of the most brilliant physicians in the US but there are drawbacks to being that intelligent. It seemed that the physicians weren't effective at communicating medical information to patients.

What it does 🚀

MyMedicAI is an AI-first web-based application that uses AssemblyAI and Streamlit to translate medical data from speech to text. That information is then broken down into medical entities

Those entities are then used to pair with 3D models and data on the medical condition to put together an interactive application to educate patients on their medical history.

Facial information is also used to make the application HIPAA compliant and offer an extra layer of security in making sure only the patient sees their data given to them by their doctor.

How we built it 🛠️

MyMedicAI could not have been possible without AssemblyAI Streamlit OpenCV Three.js Webhooks

Challenges we ran into 💣

I, like many of my fellow Hackers ran into the deadly {"error": "This account has exceeded the number of allowed streams. For more information contact support@assemblyai.com"}

Even with running json_data = json.dumps({"terminate_session": True}) I would get the same error. At times the function worked, and other times it didn't.

While attempting to make my 3D models more interactive, I somehow lost the skin for the model, which is why its fully white and hard to see. If I had more time, I would be able to figure out what went wrong and fix it, but for the sake of a MVP I left it so that I could at least have a 3D model to show.

Accomplishments that we're proud of 🏆

Although I ran into errors that I couldn't fully troubleshoot, I am confident that with a bit more time they can all be resolved. Having a working prototype for an application with the potential to impact the lives of millions and even be the future of medicine is also a plus.

What we learned

How to use speech to text and a few other AI applications that, prior to this hackathon, I had never even heard of. Big Thank You to Patrick & Andrew for all the help on Devpost and the Misra Turp for the life saving Youtube Videos