The purpose of this project is to explore the extent upto which images generated using generative models can be identified with their respective generator models. Two major techniques are used. One of them is to model this problem as a classification task and use a Convolutional Neural Network to classify the images generated from respective GANs. The other technique is to train an inverse model which tries to learn the latent space of the GAN. The image is then reconstructed and compared with the original input image. If the L2 norm of the distance between two images is less than a given threshold, then it is highly likely that the image is generated with the same Generator.
The datasets used are MNIST and CIFAR-10.
The results are documented in report.pdf