Diabetic Retinopathy Identifier

Inspiration

When I was going to eye doctor because my mom told me I had too much computer screen time, I was intrigued by the process, that of which was done on my father, of Diabetic Retinopathy.

What it does

DRI takes in two image data sets: retina scans from people without DR and retina scans of people with DR. With supervised machine learning, the computer learns the differences between both data sets. By the end of all the training, validation, and testing, the program should be able to determine if an image is a retina scan with or without DR with ~95% accuracy.

How we built it

we first downloaded a lot of photos than we used tensorflow to process the data and create the CNN.

Challenges we ran into

Getting Image Data

Accomplishments that we're proud of

Achieving nearly 90% accuracy

What we learned

The differences between those who has DR and those who don't.

What's next for Diabetic Retinopathy Identifier

Implementing the program into a USB stick so that the download process doesn't require internet.

Built With

• google-colab
• python
• tensorflow