Towards Transferable Defense Against Malicious Image Edits

TL;DR

This paper introduces TDAE, a bimodal framework combining visual and textual defenses to improve the transferability and robustness of images against malicious edits in diffusion-based systems.

Contribution

The paper proposes TDAE, a novel bimodal defense framework with FDM and DPD, enhancing cross-model immunity against malicious image manipulations.

Findings

TDAE outperforms existing methods in cross-model evaluations.

FDM effectively regularizes gradients for robustness.

DPD improves transferability through iterative embedding refinement.

Abstract

Recent approaches employing imperceptible perturbations in input images have demonstrated promising potential to counter malicious manipulations in diffusion-based image editing systems. However, existing methods suffer from limited transferability in cross-model evaluations. To address this, we propose Transferable Defense Against Malicious Image Edits (TDAE), a novel bimodal framework that enhances image immunity against malicious edits through coordinated image-text optimization. Specifically, at the visual defense level, we introduce FlatGrad Defense Mechanism (FDM), which incorporates gradient regularization into the adversarial objective. By explicitly steering the perturbations toward flat minima, FDM amplifies immune robustness against unseen editing models. For textual enhancement protection, we propose an adversarial optimization paradigm named Dynamic Prompt Defense (DPD), which periodically refines text embeddings to align the editing outcomes of immunized images with those of the original images, then updates the images under optimized embeddings. Through iterative adversarial updates to diverse embeddings, DPD enforces the generation of immunized images that seek a broader set of immunity-enhancing features, thereby achieving cross-model transferability. Extensive experimental results demonstrate that our TDAE achieves state-of-the-art performance in mitigating malicious edits under both intra- and cross-model evaluations.