Abstract
Highly realistic AI generated face swap facial imagery known as deepfakes can easily deceive human eyes and have drawn much attention. Deepfake detection models have obtained high accuracies and are still improving, but the explanation of their decisions receives little attention in current research. Explanations are important for the credibility of detection results, which is essential in serious applications like the court of law. The explanation is a hard problem, apart from the deep detection models are black boxes, the particular reason is that high-quality fakes often have no human-eye-sensitive artifacts. We call the artifacts that can be detected by models but are not human-eye-sensitive as subtle artifacts. In this work, we attempt to explain model detected face swap images to humans by proposing two simple automatic explanation methods. They enhance the original suspect image to generate its more real and more fake counterfactual versions. By visually contrasting the original suspect image with the counterfactual images, it may become easier for humans to notice some subtle artifacts. The two methods operate on pixel and color spaces respectively, they do not require extra training process and can be directly applied to any trained deepfake detection models. We also carefully design new subjective evaluation experiments to verify the effectiveness of proposed enhancement methods. Experiment results show that the color space enhancement method is more preferred by the tested subjects for explaining high-quality fake images, compared to the other pixel space method and a baseline attribution-based explanation method. The enhancement methods can be used as a toolset that helps human investigators to better notice artifacts in detected face swap images and to add weights on proofs.
This work is supported by the National Key Research and Development Program of China under Grant No. 2021YFC3320103 and the National Natural Science Foundation of China (NSFC) under Grants 61902400, U19B2038, 61972395 and a grant from Young Elite Scientists Sponsorship Program by CAST (YESS).
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
Code available at https://github.com/NiCE-X/Deepfake_explain.
References
DeepFakeDetection. https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html
DFDC top performer. https://github.com/selimsef/dfdc_deepfake_challenge
Ciftci, U.A., Demir, I., Yin, L.: FakeCatcher: detection of synthetic portrait videos using biological signals. IEEE Trans. Patt. Anal. Mach. Intell. (2020)
Dolhansky, B., et al: The DeepFake detection challenge dataset. arXiv preprint arXiv:2006.07397 (2020)
Fong, R., Patrick, M., Vedaldi, A.: Understanding deep networks via extremal perturbations and smooth masks. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 2950–2958 (2019)
Fong, R.C., Vedaldi, A.: Interpretable explanations of black boxes by meaningful perturbation. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 3429–3437 (2017)
He, Y., et al.: ForgeryNet: a versatile benchmark for comprehensive forgery analysis. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4360–4369
Jiang, L., Li, R., Wu, W., Qian, C., Loy, C.C.: DeeperForensics-1.0: a large-scale dataset for real-world face forgery detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 2889–2898
Li, L., Bao, J., Yang, H., Chen, D., Wen, F.: FaceShifter: towards high fidelity and occlusion aware face swapping. arXiv preprint arXiv:1912.13457 (2019)
Li, L., et al.: Face X-ray for more general face forgery detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 5001–5010
Li, Y., Chang, M.C., Lyu, S.: In Ictu oculi: exposing AI created fake videos by detecting eye blinking. In: 2018 IEEE International Workshop on Information Forensics and Security (WIFS), pp. 1–7. IEEE (2018)
Li, Y., Yang, X., Sun, P., Qi, H., Lyu, S.: Celeb-DF: a large-scale challenging dataset for DeepFake forensics. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 3207–3216
Malolan, B., Parekh, A., Kazi, F.: Explainable deep-fake detection using visual interpretability methods. In: 2020 3rd International Conference on Information and Computer Technologies (ICICT), pp. 289–293. IEEE (2020)
Molnar, C.: Interpretable machine learning. Lulu. com (2020)
Nguyen, H.H., Fang, F., Yamagishi, J., Echizen, I.: Multi-task learning for detecting and segmenting manipulated facial images and videos. arXiv preprint arXiv:1906.06876 (2019)
Nirkin, Y., Keller, Y., Hassner, T.: FSGAN: subject agnostic face swapping and reenactment. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 7184–7193
Peng, B., et al.: DFGC 2021: a DeepFake game competition. In: 2021 IEEE International Joint Conference on Biometrics (IJCB), pp. 1–8. https://doi.org/10.1109/IJCB52358.2021.9484387
Peng, B., Fan, H., Wang, W., Dong, J., Lyu, S.: A unified framework for high fidelity face swap and expression reenactment. IEEE Trans. Circ. Syst. Video Technol. 32, 3673–3684 (2021)
Pino, S., Carman, M.J., Bestagini, P.: What’s wrong with this video? Comparing explainers for deepfake detection. arXiv preprint arXiv:2105.05902 (2021)
Rossler, A., Cozzolino, D., Verdoliva, L., Riess, C., Thies, J., Nießner, M.: FaceForensics++: learning to detect manipulated facial images. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1–11
Selvaraju, R.R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., Batra, D.: Grad-CAM: visual explanations from deep networks via gradient-based localization. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 618–626
Woods, W., Chen, J., Teuscher, C.: Adversarial explanations for understanding image classification decisions and improved neural network robustness. Nat. Mach. Intell. 1(11), 508–516 (2019)
Zi, B., Chang, M., Chen, J., Ma, X., Jiang, Y.G.: WildDeepfake: a challenging real-world dataset for DeepFake detection. In: Proceedings of the 28th ACM International Conference on Multimedia, pp. 2382–2390 (2020)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Peng, B., Lyu, S., Wang, W., Dong, J. (2022). Counterfactual Image Enhancement for Explanation of Face Swap Deepfakes. In: Yu, S., et al. Pattern Recognition and Computer Vision. PRCV 2022. Lecture Notes in Computer Science, vol 13535. Springer, Cham. https://doi.org/10.1007/978-3-031-18910-4_40
Download citation
DOI: https://doi.org/10.1007/978-3-031-18910-4_40
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-18909-8
Online ISBN: 978-3-031-18910-4
eBook Packages: Computer ScienceComputer Science (R0)