Misleading Deep-Fake Detection with GAN Fingerprints
Authors:
- 1 May 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 59-65
Abstract
Generative adversarial networks (GANs) have made remarkable progress in synthesizing realistic-looking images that effectively outsmart even humans. Although several detection methods can recognize these deep fakes by checking for image artifacts from the generation process, multiple counterattacks have demonstrated their limitations. These attacks, however, still require certain conditions to hold, such as interacting with the detection method or adjusting the GAN directly. In this paper, we introduce a novel class of simple counterattacks that overcomes these limitations. In particular, we show that an adversary can remove indicative artifacts, the GAN fingerprint, directly from the frequency spectrum of a generated image. We explore different realizations of this removal, ranging from filtering high frequencies to more nuanced frequency-peak cleansing. We evaluate the performance of our attack with different detection methods, GAN architectures, and datasets. Our results show that an adversary can often remove GAN fingerprints and thus evade the detection of generated images.Keywords
All Related Versions
Funding Information
- Ministry of Education
This publication has 19 references indexed in Scilit:
- GANprintR: Improved Fakes and Evaluation of the State of the Art in Face Manipulation DetectionIEEE Journal of Selected Topics in Signal Processing, 2020
- Media Forensics and DeepFakes: An OverviewIEEE Journal of Selected Topics in Signal Processing, 2020
- CNN-Generated Images Are Surprisingly Easy to Spot… for Now2020
- Evading Deepfake-Image Detectors with White- and Black-Box Attacks2020
- Detecting and Simulating Artifacts in GAN Fake Images2019
- Attributing Fake Images to GANs: Learning and Analyzing GAN Fingerprints2019
- Do GANs Leave Artificial Fingerprints?2019
- Deconvolution and Checkerboard Artifacts2016
- On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance PropagationPLOS ONE, 2015
- Digital Image Forensics2013