Inspiration
We were inspired by the recent climate change march and we wanted to create an environment friendly application.
What it does
Basically, we built a mobile application which allows the user to efficiently determine the category of waste he is trying to dispose by taking a picture of it.
How we built it
We trained a convolutional neural network to classify an image as either cardboard, glass, metal, paper, plastic, or trash with the fastai library (built on PyTorch). The pre-trained model is then used to predict the classification of images received through an Android application built on ReactNative and Flask.
Challenges we ran into
Due to the duration of the event, it was difficult to train the model for multiple iterations in order to increase its accuracy, which we managed to do, although not perfectly, overnight.
What we learned
Since we are relatively new to the field of machine learning, we are satisfied to have furthered our understanding about convolutional neural networks and image classification
What's next for GoGreen
Since the duration was relatively short, in a future project, more thorough data scraping should be used in order to implement more categories of waste such as compostable elements and electronics.
Log in or sign up for Devpost to join the conversation.