Inspiration
In a time where recycling is important, it can be confusing to know what to recycle and what to not. This app can help you with this.
What it does
By simply snapping a picture of the item, it will classify what it is looking at and tell you if it is recyclable or not.
How We built it
This app was built using iOS with Swift/SwiftUI. We are running the images through a CoreML model for classification.
Challenges I ran into
To create a dataset of images we needed, it was very time-consuming. After we created it, Tensorflow/Turicreate had issues training on the data due to some of the images not being correctly formatted/save. There was no way to tell which images were corrupted, so we switched to a general image classifier due to the time constraints.
Accomplishments that I'm proud of
Making our very first machine learning app.
What We learned
We learn a lot about using Apple's new UI builder, SwiftUI, and integrating it into our project. Additionally, we learned how to collect images to create a dataset for training, and how to use Tensorflow/Turicreate to use these images for training.
What's next for RURecycling?
We hope to create a better model and add more detailed instructions on how to recycle items.
Log in or sign up for Devpost to join the conversation.