The original face has been blurred with Gaussian
blur (kernel size K = 81) to hide the user identity. Our
proposed method (Revelio) can restore the blurred face
with a higher fidelity with respect to the original identity.
Gaussian blur is widely used to blur human faces in sensitive photos before the photos are posted on the Internet. However, it is unclear to what extent the blurred faces can be restored and used to re-identify the person, especially under a high-blurring setting. In this paper, we explore this question by developing a deblurring method called Revelio. The key intuition is to leverage a generative model's memorization effect and approximate the inverse function of Gaussian blur for face restoration. Compared with existing methods, we design the deblurring process to be identity-preserving. It uses a conditional Diffusion model for preliminary face restoration and then uses an identity retrieval model to retrieve related images to further enhance fidelity. We evaluate Revelio with large public face image datasets and show that it can effectively restore blurred faces, especially under a high-blurring setting. It has a re-identification accuracy of 95.9%, outperforming existing solutions. The result suggests that Gaussian blur should not be used for face anonymization purposes. We also demonstrate the robustness of this method against mismatched Gaussian kernel sizes and functions, and test preliminary countermeasures and adaptive attacks to inspire future work.
The original clear image \(y\) is Gaussian blurred to generate image \(x\) (❶). This blurred image \(x\) is the input to Revelio. To recover the original face, we first use a Base Model \(M_B\) to perform preliminary restoration (❷). Then in step ❸, we use pre-recovered image \(y_b\) for identity retrieval. In step ❹, we use the matched identity to find more reference images to fine-tune the model. Finally, using the fine-tuned model (\(M_F\)), we can further improve the fidelity of the restored image (❺).
Because the method can potentially enable misuse (e.g., deanonymization), the code is not publicly released. Researchers may request access for legitimate academic purposes by emailing the author . Please include:
Subject: Revelio Code Access Request
To: shuow6@illinois.edu
Dear authors,
I am a researcher XXX from XXX studying XXX and would like to request
access to the Revelio codebase for academic replication purposes. I confirm that I will
use the code responsibly and will not employ it in ways that violate user privacy.
Best regards,
[Your Name]
[Institution]
We may ask for additional information before granting access.
@inproceedings{revelio2026,
author = {Zhai, Haoyu and Wang, Shuo and Naghavi, Pirouz and Hao, Qingying and Wang, Gang},
title = {Revelio: Blurred Images Can Still Disclose Your Identity},
booktitle = {2026 IEEE Symposium on Security and Privacy (SP)},
year = {2026}
}