I was inspired by the recent slew of new ChatGPT integrations, as well as the release of GPT-4. My friends and I use a bunch of social media, and I thought that this would be a funny way to make the picture taking more interesting.
What it does
Allows each user to submit a limited number of photos per day using a mobile camera interface and then scores the beauty of the user's face using a machine learning model (in the case of group photos, it averages out the scores of each face). Users are then able to view their photo and their score ranked among all other photos.
For the sake of demonstration purposes, users currently have an unlimited number of photos they can take.
How we built it
The interface was built with React Native and the backend was built with Flask and AWS. The machine learning model is largely built upon the work of the public repository https://github.com/asiryan/HowCuteAmI.
Challenges we ran into
There was initially a lot of problems with finding a good model. I ended up settling for something that was more lighweight than what I initially set out to use. There are also a lot of new things I learned about integrating Flask with AWS and React Native.
Accomplishments that we're proud of
The app is fully functional and is ready to be tested on TestFlight.
What we learned
How to use Flask in combination with AWS.
What's next for Face Score
I would love to be able to actually publish this app on the app store with some more additional features. Reinforcing the limited number of photos per day and possibly implementing a notifications-based timing event that limits the user to taking pictures at random intervals during the day is something that I feel would really boost the interactivity of this application. I would also love to revamp the UI.