Skin Disease Detection

Inspiration

What inspired me to make this project is that I wanted this project to be a starter for something larger and more impactful. A major part was that facial traits, especially on the skin, are among the most common health concerns worldwide. Having a detection tool that can help identify them earlier can make a significant impact on those who don't have easy access to professionals.

What it does

The tool works by bringing you to a local website for Streamlit, where you can take a picture of your skin or find one online to detect if it is a skin disease or not.

How we built it

I built it with the help of Chat GPT and Claude AI in Python, as well as datasets from Hugging Face.

Challenges we ran into

The major challenge that I ran into was that the specific dataset I was using required me to download the full 1 million indexes. I had tried to only extract the first thousands, but none of that worked, and I ran into errors with other datasets. As a result, the project I'm submitting is a model that has not been trained with its data to identify the specific diseases, hence why it only gives a percentage of whether it is or not a disease.

Accomplishments that we're proud of

Accomplishments that I'm proud of are learning to use datasets from Hugging Face, sufficiently training models based on the data, and understanding a good portion of the concept of training models with data and images.

What we learned

In the future, I should do some research on specific datasets and their requirements. I should also make sure that I have a good dataset for a specific-sized project.