Inspiration

DermoAI was inspired by the everyday struggle that many people deal with skin conditions. Although it is our largest organ, there seems to be limited care/lack of knowledge for it in many cases. Whether it be long wait times for dermatologists, limited access to care, or just not knowing what's wrong with their skin.

# What it does

The idea is simple, but effective: Take a picture and allow AI to analyze it. DermoAI analyzes the skin, identifies the most likely condition, its severity, and recommends immediate treatment actions and further steps to be taken. If the condition is severe enough, it also gives you the ability to locate your nearest dermatologist to schedule an appointment.

How we built it

We used Python, JavaScript, HTML, and the Hugging Face model to build a cohesive website for DermoAI. We also used React, Claude, and Google Maps to develop the UI and allow users to find their local dermatologists. The backend runs on Node.js with Express, deployed on Vercel and Render.

Challenges we ran into

Accuracy was the biggest issue we ran into. When going through different data sets, we kept on getting the wrong skin condition detections through the AI whenever we were running the sample images. We also spent hours debugging deployment issues and API integrations.

Accomplishments that we're proud of

Our biggest accomplishment is getting this website to work in general. But we're also really proud of how accessible the UI is and how we developed a working prototype that can analyze skin conditions in a matter of seconds.

What we learned

We learned about dermatology, especially the skin conditions and the issues that arise from them. We also learned the importance of responsible AI deployment in healthcare contexts.

What's next for DermoAI

We would want to build a native app for DermoAI to make it even more accessible to people. We're also planning to add condition tracking and expand to more skin conditions.