After I generated the images to train the model, I used a convolutional neural network to learn on those images. A limitation for my project was that I had relatively little data, so this confusion matrix, which shows correct predictions in the top left and bottom right corners, and incorrect predictions in the bottom left and the top right, is based on the small amount of test data that I had. Nonetheless, it has an accuracy level of (correct predictions / total predictions ) =~ 83%, so it is promising and should be trained and tested with larger datasets.