In this project a DCGAN will be implemented for the purpose of generation low resolution faces. It is planned to gradually further develop this model to approximate the U-Net Based Discriminator for Generative Adversarial Networks of Schönfeld et. al..