MetaCloak-JPEG: JPEG-Robust Adversarial Perturbation for Preventing Unauthorized DreamBooth-Based Deepfake Generation
Tanjim Rahaman Fardin, S M Zunaid Alam, Mahadi Hasan Fahim, Md Faysal Mahfuz

TL;DR
MetaCloak-JPEG introduces a differentiable JPEG layer to enhance adversarial image protection against deepfake models, effectively disrupting fine-tuning even after JPEG compression.
Contribution
It proposes a novel Differentiable JPEG layer using STE within a meta-learning framework to improve adversarial robustness against JPEG-processed images.
Findings
Achieves 91.3% JPEG survival rate under attack.
Outperforms existing methods across all tested JPEG quality factors.
Maintains high image quality with 32.7 dB PSNR.
Abstract
The rapid progress of subject-driven text-to-image synthesis, and in particular DreamBooth, has enabled a consent-free deepfake pipeline: an adversary needs only 4-8 publicly available face images to fine-tune a personalized diffusion model and produce photorealistic harmful content. Current adversarial face-protection systems -- PhotoGuard, Anti-DreamBooth, and MetaCloak -- perturb user images to disrupt surrogate fine-tuning, but all share a structural blindness: none backpropagates gradients through the JPEG compression pipeline that every major social-media platform applies before adversary access. Because JPEG quantization relies on round(), whose derivative is zero almost everywhere, adversarial energy concentrates in high-frequency DCT bands that JPEG discards, eliminating 60-80% of the protective signal. We introduce MetaCloak-JPEG, which closes this gap by inserting a…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
