Betwixt
The Project
Betwixt is an AI-powered hardware solution that uses image/video classification models, Firebase, and Python 3 on the Raspberry Pi to digitally categorize and physically sort waste into its correct category.
Team Comments
I chose to make this project because...When my city started implementing a three-waste program---a waste bin with three categories: organics, recyclables, and regular trash---I struggled and often miscategorized waste. The need to educate others and reduce the labor, time, and financial costs, to fix misplaced waste inspired this project
What I found difficult and how I worked it outA difficult part was getting categorization results from the Android device to control the hardware. I researched into mini-computers that I could use to run Python 3 scripts. My computer science teacher lended me a Raspberry Pi, and after setting it up, I could connect the two devices together.
Next time, I would...I would definitely train the image classification model within the context of waste separation/categorization. The model I used from App Inventor was trained on more general objects rather than waste. I hope to train and add a better use-specific model.
