We initially started off with the unoriginal idea to create a reminder application. Although there were many out there to use, we figured it would be a good place for a few novice programmers to get their feet wet at a hackathon. However, we were encouraged by our colleagues to aim higher and build something meaningful.
After some brainstorming, we decided to build something that helped others and focused on small basal carcinoma. The program utilizes trained IBM visual recognition models and analyzes submitted photos of irritated skin.
A few challenges we ran into were familiarizing ourselves with machine learning especially since it was our first time working with it. Another issue was developing a user-friendly front end that joined the user submissions and the IBM cloud.
An accomplishment that we are proud of is working with new technologies that pushed us outside of our comfort zone. The other is making something meaningful and realizing that the effort we put into our code could make a difference in the world.
Technically we learned about how machine learning works at a fundamental level and gained experience with FLASK which allowed us to connect our Python API calls with our web front end. More generally, we learned how to join different technologies in order to produce a final project.
We plan to take dermView to the next level by increasing the sample size of our data sets and build using submissions from users. We also plan to learn more machine learning technologies in order to create a more robust, accurate, and scalable platform for multiple diagnoses.