SkinACure | Devpost

Inspiration

Having a friend who suffers from a severe skin disease and having studied many skin diseases during HOSA, our team was interested in exploring a solution towards automating the treatment and diagnosis of various skin diseases.

What it does

We created a 3-step system that first identifies the patient and diagnosis the patient with the according skin disease. From there, the system communicates to our hardware side that picks out the medication for the respective skin disease. The system then dispenses the medication into a wearable skin patch that uses iontophoresis technology for transdermal drug administration.

How we built it

We used HTMl, CSS, JS for the front-end of our website. To scan and detect different types of diseases we first trained and built our model on Google Colab. Then deployed it through with webcam. The AI model detects different skin diseases and communicates to the hardware that uses an Arduino to dispense the medication into a funnel that is applied on the skin patch. The skin patch uses iontophoresis technology.

Challenges we ran into

Training the model Connecting the software side to send information to the hardware side Building and finding the optimal places to place hardware parts

Accomplishments that we're proud of

Implementing an AI Model and having it connect to a hardware component. In general, we are really happy to be finished our project.

What we learned

We learnt how to train and deploy deep learning models in python with PyTorch. We also learn to use the ChatGpt API. We picked up some knowledge in the hardware component as well.