Marius Stokkedal & Sebastian Bengtsson
This project intends to result in a defense for an image classifier of aerial photos - in this case, a Convolutional neural network (CNN) model. The attack model used is the Fast Gradient Sign Method (FGSM) and the models that are used for defense are Support Vector Machine (SVM), k-nearest Neighbors (KNN), and Decision Tree (DT). The first attempt at defense models does not work as hoped - most of the data is classified as being under attack regardless of whether it is or not.