Detecting GAN generated Fake Images using Co-occurrence Matrices

• Published in: Electronic Imaging Vol. 31; no. 5; pp. 532-1 - 532-7
• Main Authors: Nataraj, Lakshmanan, Mohammed, Tajuddin Manhar, Manjunath, B. S., Chandrasekaran, Shivkumar, Flenner, Arjuna, Bappy, Jawadul H., Roy-Chowdhury, Amit K.
• Format: Journal Article
• Language: English
• Published: 7003 Kilworth Lane, Springfield, VA 22151 USA Society for Imaging Science and Technology 13.01.2019
• Subjects: Deep Learning; Fake Images; Fake News; Gan Image Detection; Image Forensics; Multimedia Security
• ISSN: 2470-1173; 2470-1173
• DOI: 10.2352/ISSN.2470-1173.2019.5.MWSF-532

The advent of Generative Adversarial Networks (GANs) has brought about completely novel ways of transforming and manipulating pixels in digital images. GAN based techniques such as Image-to-Image translations, DeepFakes, and other automated methods have become increasingly popular in creating fake images. In this paper, we propose a novel approach to detect GAN generated fake images using a combination of co-occurrence matrices and deep learning. We extract co-occurrence matrices on three color channels in the pixel domain and train a model using a deep convolutional neural network (CNN) framework. Experimental results on two diverse and challenging GAN datasets comprising more than 56,000 images based on unpaired image-to-image translations (cycleGAN [1]) and facial attributes/expressions (StarGAN [2]) show that our approach is promising and achieves more than 99% classification accuracy in both datasets. Further, our approach also generalizes well and achieves good results when trained on one dataset and tested on the other.