Fourier Spectrum Discrepancies in Deep Network Generated Images

11/15/2019

∙

by   Tarik Dzanic, et al.

∙

21

∙

share

Advancements in deep generative models such as generative adversarial networks and variational autoencoders have resulted in the ability to generate realistic images that are visually indistinguishable from real images. In this paper, we present an analysis of the high-frequency Fourier modes of real and deep network generated images and the effects of resolution and image compression on these modes. Using this, we propose a detection method based on the frequency spectrum of the images which is able to achieve an accuracy of up to 99.2% in classifying real, Style-GAN generated, and VQ-VAE2 generated images on a dataset of 2000 images with less than 10% training data. Furthermore, we suggest a method for modifying the high-frequency attributes of deep network generated images to mimic real images.