Cows and methane. Least to say, our team was greatly excited when we saw the sustainability topic! Carbon capture, geoengineering, all the fun stuff. But most personal to us, was that we weren't doing anything ourselves even though we were highly interested. What did we need?! We need a Tony Stark assistant. We need an object detection algorithm, where when we take a photo of what we picked up, we get bonus points (depending on the object's carbon footprint and its size).
What it does
With an object detection algorithm, take what people picked up and report it on a map (so people know where the litter is)! But even more importantly, take those objects understand their importance and size (by bounding boxes), and create boost points! Then your boost points will be added to the leaderboard and you are ranked!
How we built it
Split into the machine learning part vs the mobile development centered around that algorithm. Both parts were developed at the same time, and plugged in into the last moment.
Challenges we ran into
Tons of Flutter dependencies going haywire (and 24 hours of resolving dependencies). Pretty fun and all! It was the dependency influenza.
Accomplishments that we're proud of
Getting the AI model to a mobile version where it could run on CPU... wow it took long but it was worth it (the core feature of our app)!
What we learned
Learned about tflite, learned about how to handle when firebase crashes. Learned about gamification and how to execute on the aspect with the best UX. And we learned how to have a lot of fun while simultaneously being hungry and frustrated!
What's next for EcoSnap
Reporting for organizations and cities and more on how cities can interact with these locations, interact with these individuals that are hyperfocused on cleaning up litter in certain areas! A data analytic part for these cities would also be great!