This project aims to compare the results of Variational Autoencoders (VAE), Deep Convolutional Generative Adversarial Networks (DCGAN) and Capsule Generative Adversarial Networks (CapsuleGAN) in generating images of handwritten digits from the MNIST dataset. Their performance is evaluated via loss computation and the generated images are compared after a certain fixed number of epochs.
The link to the project approach and results generated can be found here: Paper link