Inspiration
At Benson there are always three -- very intimidating -- bins, compost, trash, and recycling. This can make you doubt if you actually know how to sort your trash which gave us the idea to make our own convolutional neural network to sort trash!
What it does
This Flutter mobile app scans your trash, runs it through a CNN built in python hosted by flask, then lets you know which bin it thinks you should throw your trash in along with its confidence. There is also a feature to view trash bins near you on a map, and then thanks to our backend made with Node.js you can upload trash cans to the MongoDB collection. The community will help us keep our streets clean by letting us know where trash cans are!
How we built it
We built this mainly using Dart code because it is a flutter app, but we also had to use python for the CNN and JS for connecting to MongoDB. We wrote API endpoints, hosted flask servers, wrote logic, created UI elements in flutter, and trained a convolutional neural network on 22.5 thousands photos!
Challenges we ran into
Getting and posting coordinate to mongoDB proved to be much more complicated than anticipated. Also, navigating flutters powerful UI elements was quite a steep learning curve.
We got the map working a little bit after filming the demo video. Here is another video of just the map working https://youtube.com/shorts/1iCzrsz9V7Y?feature=share
Accomplishments that we're proud of and what we learned
We are proud of all our features! Everything in this project was new to us, but we managed to learn new programs, train a CNN, and connect backend to frontend in multiple ways!
What's next for Trash Classification AI Mobile App
Launching our app would be great for the community! It would teach people how to properly throw away trash and allow people to find nearby trash cans which would be sure to clean up our streets from litter!
Log in or sign up for Devpost to join the conversation.