Author(s): Aisha Saleem, Narangerel Boldbaatar, Nicolas Decavel-Bueff, Qianti Min
We want to use Machine Learning techniques, specifically Classification, to raise awareness in regards to what can and what should not be recycled. The final product will include a model that, given an image, will identify whether or not the main item in the picture is Recyclable (Plastic, Glass, Metal, Cardboard, Paper), E-Waste, or Trash. Later, we will also deploy this model in a web app that will incorporate user feedback to give us an idea of the model's "use-case" accuracy while populating more images into our training dataset.
To run our model, simply open up 'Training_Model.ipnyb' in a juypter notebook and the rest should be self-explanatory from the code. Our web-deployed app can be seen at: https://recycling-test.onrender.com/.