Towards Robust Protective Perturbation against DeepFake Face Swapping
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914186289741824 |
|---|---|
| author | Yao, Hengyang Li, Lin Sun, Ke Qiu, Jianing Chen, Huiping |
| author_facet | Yao, Hengyang Li, Lin Sun, Ke Qiu, Jianing Chen, Huiping |
| contents | DeepFake face swapping enables highly realistic identity forgeries, posing serious privacy and security risks. A common defence embeds invisible perturbations into images, but these are fragile and often destroyed by basic transformations such as compression or resizing. In this paper, we first conduct a systematic analysis of 30 transformations across six categories and show that protection robustness is highly sensitive to the choice of training transformations, making the standard Expectation over Transformation (EOT) with uniform sampling fundamentally suboptimal. Motivated by this, we propose Expectation Over Learned distribution of Transformation (EOLT), the framework to treat transformation distribution as a learnable component rather than a fixed design choice. Specifically, EOLT employs a policy network that learns to automatically prioritize critical transformations and adaptively generate instance-specific perturbations via reinforcement learning, enabling explicit modeling of defensive bottlenecks while maintaining broad transferability. Extensive experiments demonstrate that our method achieves substantial improvements over state-of-the-art approaches, with 26% higher average robustness and up to 30% gains on challenging transformation categories. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_07228 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Towards Robust Protective Perturbation against DeepFake Face Swapping Yao, Hengyang Li, Lin Sun, Ke Qiu, Jianing Chen, Huiping Computer Vision and Pattern Recognition Artificial Intelligence Cryptography and Security Machine Learning DeepFake face swapping enables highly realistic identity forgeries, posing serious privacy and security risks. A common defence embeds invisible perturbations into images, but these are fragile and often destroyed by basic transformations such as compression or resizing. In this paper, we first conduct a systematic analysis of 30 transformations across six categories and show that protection robustness is highly sensitive to the choice of training transformations, making the standard Expectation over Transformation (EOT) with uniform sampling fundamentally suboptimal. Motivated by this, we propose Expectation Over Learned distribution of Transformation (EOLT), the framework to treat transformation distribution as a learnable component rather than a fixed design choice. Specifically, EOLT employs a policy network that learns to automatically prioritize critical transformations and adaptively generate instance-specific perturbations via reinforcement learning, enabling explicit modeling of defensive bottlenecks while maintaining broad transferability. Extensive experiments demonstrate that our method achieves substantial improvements over state-of-the-art approaches, with 26% higher average robustness and up to 30% gains on challenging transformation categories. |
| title | Towards Robust Protective Perturbation against DeepFake Face Swapping |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Cryptography and Security Machine Learning |
| url | https://arxiv.org/abs/2512.07228 |