Imagine Cup Diabetes Detection System.
What it does
Detects whether a person has symptoms of hyperthyroidism, high cholesterol or diabetes with a simple photo of the eye.
How we built it
We used over 2,000 images to train the system to classify the different systems.
Challenges we ran into
Not enough images on the internet related to a specific disease
Accomplishments that we're proud of
Being able to detect the disease with 86% accuracy and detect that my mom had hyperthyroidism with accuracy so that maybe we can move forward to tests. Integration of systems and unit tests.
What we learned
System Architecture Machine Learning Using CV and GAN Service Communication in real-time with the app Generating Analytics from Results
What's next for 4.dreye
Continue to improve the accuracy of the system and work on being able to detect diseases within the retina.