The aim of this project is to process pictures of one piece of garbage and classify it into 5 categories consisting of glass, paper, metal, plastic, and cardboard. We have additionally produced a dataset that contains around 400-500 pictures for every category, that was each hand collected and gathered from online sources . A convolutional neural network is employed during this project, created on Google’s open source code library for dataflow pro- gramming - Tensorflow. This project additionally uses Transfer Learning to use already existing predicting models of image classification and rewire them for finishing the task at hand