Transferable Adversarial Attack on Image Tampering Localization

TL;DR

This paper introduces an adversarial attack method that tests the robustness of image tampering localization algorithms, demonstrating significant accuracy reduction while maintaining image quality.

Contribution

It presents a novel transferability-based adversarial attack scheme specifically designed for evaluating image tampering localization methods.

Findings

Attack significantly reduces localization accuracy

High visual quality of attacked images is preserved

Effective in both white-box and black-box scenarios

Abstract

It is significant to evaluate the security of existing digital image tampering localization algorithms in real-world applications. In this paper, we propose an adversarial attack scheme to reveal the reliability of such tampering localizers, which would be fooled and fail to predict altered regions correctly. Specifically, the adversarial examples based on optimization and gradient are implemented for white/black-box attacks. Correspondingly, the adversarial example is optimized via reverse gradient propagation, and the perturbation is added adaptively in the direction of gradient rising. The black-box attack is achieved by relying on the transferability of such adversarial examples to different localizers. Extensive evaluations verify that the proposed attack sharply reduces the localization accuracy while preserving high visual quality of the attacked images.