Attributing Fake Images to GANs: Learning and Analyzing GAN Fingerprints

Yu, Ning; Davis, Larry S.; Fritz, Mario

doi:10.1109/iccv.2019.00765

Cited by 374 publications

(281 citation statements)

References 55 publications

Supporting

Mentioning

274

Contrasting

Unclassified

Order By: Relevance

“…This observation suggests the presence of different artifacts peculiar of each specific GAN model. Recently, it has also been shown [22] that a deep network can reliably discriminate images generated with different architectures. However, the network requires intensive training on an aligned dataset, and there is no hint, let alone exploitation, of the presence of GANinduced fingerprints.…”

Section: Related Workmentioning

confidence: 99%

Do GANs Leave Artificial Fingerprints?

Marra

Gragnaniello

Verdoliva

et al. 2019

2019 IEEE Conference on Multimedia Information Processing and Retrieval (MIPR)

267

177

View full text Add to dashboard Cite

In the last few years, generative adversarial networks (GAN) have shown tremendous potential for a number of applications in computer vision and related fields. With the current pace of progress, it is a sure bet they will soon be able to generate high-quality images and videos, virtually indistinguishable from real ones. Unfortunately, realistic GAN-generated images pose serious threats to security, to begin with a possible flood of fake multimedia, and multimedia forensic countermeasures are in urgent need. In this work, we show that each GAN leaves its specific fingerprint in the images it generates, just like real-world cameras mark acquired images with traces of their photo-response non-uniformity pattern. Source identification experiments with several popular GANs show such fingerprints to represent a precious asset for forensic analyses.

show abstract

Section: Related Workmentioning

confidence: 99%

Do GANs Leave Artificial Fingerprints?

Marra

Gragnaniello

Verdoliva

et al. 2019

2019 IEEE Conference on Multimedia Information Processing and Retrieval (MIPR)

267

177

View full text Add to dashboard Cite

show abstract

“…A typical way to design a real vs. GAN fake image classifier is to collect a large number of GAN generated images from one or multiple pre-trained GAN models and train a binary classifier [4], [5]. Unfortunately, in real world applications, we generally have no access to the specific model used by the attacker.…”

Section: Introductionmentioning

confidence: 99%

Detecting and Simulating Artifacts in GAN Fake Images

Karaman

Chang

2019

2019 IEEE International Workshop on Information Forensics and Security (WIFS)

398

246

View full text Add to dashboard Cite

To detect GAN generated images, conventional supervised machine learning algorithms require collection of a number of real and fake images from the targeted GAN model. However, the specific model used by the attacker is often unavailable. To address this, we propose a GAN simulator, AutoGAN, which can simulate the artifacts produced by the common pipeline shared by several popular GAN models. Additionally, we identify a unique artifact caused by the up-sampling component included in the common GAN pipeline. We show theoretically such artifacts are manifested as replications of spectra in the frequency domain and thus propose a classifier model based on the spectrum input, rather than the pixel input. By using the simulated images to train a spectrum based classifier, even without seeing the fake images produced by the targeted GAN model during training, our approach achieves state-of-the-art performances on detecting fake images generated by popular GAN models such as CycleGAN.

show abstract

“…Thus far, most state-of-the-art fake image detection methods are proposed based on convolutional neural network (CNN), which roughly fall into three categories by their input feature types, i.e., image-based methods [4,37,38,54], fingerprint-based methods [57], and spectrum-based methods [14,59].…”

Section: Introductionmentioning

confidence: 99%

“…These three types of methods [14,54,57] all demonstrate their usefulness, in achieving the state-of-the-art performance for GANsynthesized fake image detection.…”

Section: Introductionmentioning

confidence: 99%

FakePolisher: Making DeepFakes More Detection-Evasive by Shallow Reconstruction

Huang

Juefei-Xu

Wang

et al. 2020

Proceedings of the 28th ACM International Conference on Multimedia

View full text Add to dashboard Cite

At this moment, GAN-based image generation methods are still imperfect, whose upsampling design has limitations in leaving some certain artifact patterns in the synthesized image. Such artifact patterns can be easily exploited (by recent methods) for difference detection of real and GAN-synthesized images. However, the existing detection methods put much emphasis on the artifact patterns, which can become futile if such artifact patterns were reduced. Towards reducing the artifacts in the synthesized images, in this paper, we devise a simple yet powerful approach termed FakePolisher that performs shallow reconstruction of fake images through a learned linear dictionary, intending to effectively and efficiently reduce the artifacts introduced during image synthesis. In particular, we first train a dictionary model to capture the patterns of real images. Based on this dictionary, we seek the representation of DeepFake images in a low dimensional subspace through linear projection or sparse coding. Then, we are able to perform shallow reconstruction of the 'fake-free' version of the DeepFake image, which largely reduces the artifact patterns DeepFake introduces. The comprehensive evaluation on 3 state-of-the-art DeepFake detection methods and fake images generated by 16 popular GAN-based fake image generation techniques, demonstrates the effectiveness of our technique. Overall, through reducing artifact patterns, our technique significantly reduces the accuracy of the 3 state-of-theart fake image detection methods, i.e., 47% on average and up to 93% in the worst case. Our results confirm the limitation of current fake detection methods and calls the attention of DeepFake researchers and practitioners for more general-purpose fake detection techniques. CCS CONCEPTS • Security and privacy → Human and societal aspects of security and privacy; • Computing methodologies → Computer vision.

show abstract

Attributing Fake Images to GANs: Learning and Analyzing GAN Fingerprints

Cited by 374 publications

References 55 publications

Do GANs Leave Artificial Fingerprints?

Do GANs Leave Artificial Fingerprints?

Detecting and Simulating Artifacts in GAN Fake Images

FakePolisher: Making DeepFakes More Detection-Evasive by Shallow Reconstruction

Contact Info

Product

Resources

About