Abstract
In this paper, we investigate the impact of adversarial attacks on identity encoders within a realistic de-identification framework. Our experiments show that the transferability of attacks transfers from an external surrogate model to the system model (e.g., CosFace to ArcFace) allows the adversary to cause identity information to leak in a sufficiently sensitive face recognition system. We present experimental evidence and propose strategies to mitigate this vulnerability. Specifically, we show how fine-tuning on adversarial examples helps to mitigate this effect for distortion-based attacks (i.e., snow, fog, etc.), while a simple low-pass filter can attenuate the effect of adversarial noise without affecting the de-identified images. Our mitigation results in a de-identification system that preserves its functionality while being significantly more robust to adversarial noise.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Rosberg, F., Englund, C., Aksoy, E. E., & Alonso-Fernandez, F. (2026). Adversarial Attacks and Identity Leakage in De-Identification Systems: An Empirical Study. https://omanscience.com/en/articles/adversarial-attacks-and-identity-leakage-in-de-identification-systems-an-empirical-study
MLA 9
Rosberg, Felix, et al. "Adversarial Attacks and Identity Leakage in De-Identification Systems: An Empirical Study." https://omanscience.com/en/articles/adversarial-attacks-and-identity-leakage-in-de-identification-systems-an-empirical-study.
Chicago (author–date)
Rosberg, Felix, Cristofer Englund, Eren Erdal Aksoy, and Fernando Alonso-Fernandez. 2026. "Adversarial Attacks and Identity Leakage in De-Identification Systems: An Empirical Study." https://omanscience.com/en/articles/adversarial-attacks-and-identity-leakage-in-de-identification-systems-an-empirical-study.
Harvard
Rosberg, F., Englund, C., Aksoy, E. E. and Alonso-Fernandez, F. (2026) 'Adversarial Attacks and Identity Leakage in De-Identification Systems: An Empirical Study', Available at: https://omanscience.com/en/articles/adversarial-attacks-and-identity-leakage-in-de-identification-systems-an-empirical-study.
Vancouver
Rosberg F, Englund C, Aksoy EE, Alonso-Fernandez F. Adversarial Attacks and Identity Leakage in De-Identification Systems: An Empirical Study. https://omanscience.com/en/articles/adversarial-attacks-and-identity-leakage-in-de-identification-systems-an-empirical-study
IEEE
F. Rosberg, C. Englund, E. E. Aksoy, and F. Alonso-Fernandez, "Adversarial Attacks and Identity Leakage in De-Identification Systems: An Empirical Study," https://omanscience.com/en/articles/adversarial-attacks-and-identity-leakage-in-de-identification-systems-an-empirical-study.