Inspiration
On our school campus, we notice lots and lots of trash, but nobody seems bothered!
What it does
Our app gets image input of your trash, either through live feed or through image upload, and then it classifies it into a bin (recycling bin, hazard bin, etc.) and gives you points based on its reusability.
How we built it
We first created a dataset containing thousands of images and trained a CNN image classifier. After yielding unideal results, we decided to borrow from a multitude of datasets and train a YOLO model to classify the waste based on our specific classes.
Challenges we ran into
Training the model, creating the dataset, and having a fully-functional backend.
Accomplishments that we're proud of
Training the model, creating the dataset, and having a fully-functional backend.
What we learned
We learned how to do image classification as well as image classification in real time!
What's next for Trash Talker
More gamification to incentivize users to clean our environment more.
Log in or sign up for Devpost to join the conversation.