Recycler

Inspiration

RecycleMania is happening currently and the idea kind of sprung up from that.

What it does

It uses Neural Networks to make a simple classifier that classifies waste into 6 different categories using the dataset in the link: https://github.com/garythung/trashnet/blob/master/data/dataset-resized.zip

How we built it

We used Tensorflow, Keras, Flask, and Java to build a simple prototype to classify waste.

Challenges we ran into

Models not working as expected, Wiring up the APIs, Exporting the trained model.

Accomplishments that we're proud of

Trained a neural network to be pretty decent at classifying 6 categories.

What we learned

Challenges of training a deep neural network, to never forget to standardize variables, and the frustration and the fun of making something that seemingly learns on its own.

What's next for Recycler

Use a better data set to create a better classifier and maybe use Arduino and tensorflow-lite to make something like a 'smart-bin' that can sort recyclable and non-recyclable waste on its own.