Inspiration
Crossover was inspired by Google's DeepDream algorithm, which generates dream-like images using machine learning:
Crossover's approach allows for a more easily accessible (and fun) way to transplant elements of filters into your own images.
What it does
Crossover uses Style Transfer's custom machine learning algorithms to mix user photos with filters.
How we built it
We utilized Expo's SDK and React Native in order to create a front-end app for users to take photos and apply custom styles to them.
Challenges we ran into
Getting started with React Native was challenging as we had little experience with mobile app development and React.
Accomplishments that we're proud of
Our biggest accomplishment was completing a total 180 and switching ideas with 16 hours left in PennApps. Additionally, maintaining server uptime after multiple power outages due to the hurricane (our servers are based in Miami) was pretty impressive.
What we learned
Our entire team is now fairly familiar with technologies like React, React Native, and Expo, each of which we had 0 experience with in the past.
What's next for Crossover
Our next major step is increasing server performance by caching filters in order to create live video feedback in the camera mode, similar to Snapchat's Lenses.
Log in or sign up for Devpost to join the conversation.