I implemented a Deep Convolutional GAN. This DCGAN was made of a pair of multi-layer neural networks that compete against each other until one learns to generate realistic images of faces.
Those images below had been generated by the DCGAN that I built.