The Figure 3 below shows that GAN comprises of two main parts: generator and discriminator. As the name suggests, generator is responsible to generate new (fake) samples while discriminator attempts to distinguish the real and fake ones. The main objective of training a GAN is to make the generator able to generate new samples such that those samples are indistinguishable by the discriminator. Once this happens, it means that our generator is now able to create samples which the quality is already as good as the originals. As I've mentioned earlier, we are going to work on two-dimensional data since it is a lot simpler as compared to the MNIST dataset we saw earlier.

The use of accurate scanning transmission electron microscopy (STEM) image simulation methods require large computation times that can make their use infeasible for the simulation of many images. Other simulation methods based on linear imaging models, such as the convolution method, are much faster but are too inaccurate to be used in application. In this paper, we explore deep learning models that attempt to translate a STEM image produced by the convolution method to a prediction of the high accuracy multislice image. We then compare our results to those of regression methods. We find that using the deep learning model Generative Adversarial Network (GAN) provides us with the best results and performs at a similar accuracy level to previous regression models on the same dataset.