The SkinSense App

Inspiration

When deciding on a topic to build an app for this competition, we came across the issue of Melanoma, something one of our teammates had dealt with in the past. We recognized that if detected early enough, Melanoma was treatable and most of its harms could be mitigated, but sadly it's really hard and costly to detect Melanoma fast. That's why we decided to build an app to do just that.

What it does

Our app uses a Machine Learning model to take any picture and output a percentage of how likely the person in the picture has Melanoma based on it.

How we built it

We built the Model using Python and specifically used the TensorFlow libraries. Additionally, we used JavaScript to build the GUI.

Challenges we ran into

One challenge we ran into was training the Model and the computing required to do so. To combat that, we decided to use Google CoLab and the GPUs that they provide.

Accomplishments that we're proud of

Our model ended with a 91% accuracy, which is extremely good in the world of disease detection using machines.

What we learned

Through building this app, we learned how to use various Python libraries such as Tensorflow and Keras and we also learned how to build a GUI in JavaScript using libraries such as React, React Native, and React Navigation.

What's next for The SkinSense App

Our plan further is to market it to as many people as possible and spread awareness of technology like this to those who need it. We plan to partner with the right organizations, talk to the right officials, and also gain recognition through the RSA competition.

Built With

• expo.io
• java
• keras
• python
• react
• react-native
• react-navigation
• tensorflow