Segragating trash is prone to human error and is a thankless activity. We realised that we could easily solve this by taking the decision making out of the user's hand and teaching a neural network to identify and segragate the trash.
What it does
It aut segregates trash
How we built it
Using RaspberryPi and neural networks with some custom hardware hacking
Challenges we ran into
Marrying hardware and software is always a challenge. To get it all working together did take us some time.
Accomplishments that we're proud of
It works! The trash can successfully recognises and autotrashes over 1000 objects approproately between recycle and compost
What we learned
What's next for AutoTrash - Cerberus
Make a production ready design and go to market strategy. Also work on retraining the model via collective data of all trash cans.