Radpi

Inspiration

We were inspired by a foundation trying to get healthcare support in Eswatini.

What it does

It takes x-rays, images, and scans, identifying issues that are incredibly hard to see with the human eye.

How we built it

We trained yolo and other classification models to identify issues like burns, and fractures that may be hard to see, then using node.js, and three.js to build the website. We also used a raspberry pi to process the information, and this way it doesn't need any WiFi

Challenges we ran into

It was hard to get the raspberry pi to function off of an unstable power source like a power bank because it seemed to crash every time we used out built in llm.

Accomplishments that we're proud of

We're proud of how accurate the models are able to detect fractures, and conditions that are nearly impossible for us to see on our own.

What we learned

We learn

What's next for Radpi

We'd like to add more models and make the whole process more accurate. Perhaps, trying the same thing with a more powerful raspberry pi.

Built With

• node.js
• python
• qwen
• three.js
• yolo