Glaucoma Gone

• 

  Diagram of Pipeline

Inspiration

My inspiration was Invisalign which has an app that scans your teeth to see if your teeth are starting to straighten up.

What it does

My app takes a picture of a person's eye and uses artificial intelligence to see if you have Glaucoma. The results will be more accurate if the user attaches a condenser lens to their phone camera so they can see the Fundus of the eye more clearly.

How we built it

We used android studio and Kotlin to create the UI. We used TensorFlow for the artificial intelligence portion. We used transfer learning to train the TensorFlow model on the MobileNetV3 model. We used some code from Yuval Altman who trained his model with EfficientNet instead of MobileNetV3. We also used some sample code to help us use Tensorflow lite to call the model from our application.

Challenges we ran into

One challenge we ran into was that our Tensorflow model's input type was different from the default so we had to edit the model's metadata.

Accomplishments that we're proud of

We are proud of the speed and accuracy of our app.

What we learned

We learned how to edit model metadata, TensorFlow lite, and transfer learning techniques.

What's next for Glaucoma Gone

Some possible steps are improving the UI and adding support for more diseases.

Built With

• android
• android-studio
• kotlin
• tensorflow