Hiding Local Manipulations on SAR Images: a Counter-Forensic Attack

TL;DR

It is demonstrated that an expert practitioner can exploit the complex nature of SAR data to obscure any signs of manipulation within a locally altered amplitude image and assess the effectiveness of the proposed counter-forensic approach across diverse scenarios, examining various manipulation operations.

Abstract

The vast accessibility of Synthetic Aperture Radar (SAR) images through online portals has propelled the research across various fields. This widespread use and easy availability have unfortunately made SAR data susceptible to malicious alterations, such as local editing applied to the images for inserting or covering the presence of sensitive targets. Vulnerability is further emphasized by the fact that most SAR products, despite their original complex nature, are often released as amplitude-only information, allowing even inexperienced attackers to edit and easily alter the pixel content. To contrast malicious manipulations, in the last years the forensic community has begun to dig into the SAR manipulation issue, proposing detectors that effectively localize the tampering traces in amplitude images. Nonetheless, in this paper we demonstrate that an expert practitioner can exploit the complex nature of SAR data to obscure any signs of manipulation within a locally altered amplitude image. We refer to this approach as a counter-forensic attack. To achieve the concealment of manipulation traces, the attacker can simulate a re-acquisition of the manipulated scene by the SAR system that initially generated the pristine image. In doing so, the attacker can obscure any evidence of manipulation, making it appear as if the image was legitimately produced by the system. This attack has unique features that make it both highly generalizable and relatively easy to apply. First, it is a black-box attack, meaning it is not designed to deceive a specific forensic detector. Furthermore, it does not require a training phase and is not based on adversarial operations. We assess the effectiveness of the proposed counter-forensic approach across diverse scenarios, examining various manipulation operations.

Figures (9)

• Figure 1: High-level description of the paper goal. Given a manipulated SAR amplitude image (left), state-of-the-art SAR forensic detectors estimate a real-valued heatmap (second column) that highlights local image inconsistencies. Our proposed counter-forensic attack simulates a SAR re-acquisition of the manipulated image. The attack leaves untouched the semantic content of the image (third column), but it conceals its tampering traces, such that the estimated heatmap (right) does not allow anymore to localize the forgery. The manipulation area is surrounded by the red contour.
• Figure 2: Example of 2D SLC SAR image $\mathbf{I}e^{j \mathbf{\Phi}}$. First, we report the amplitude component $\mathbf{I}$; then, the amplitude of the of the complex image $F(\mathbf{I}e^{j \mathbf{\Phi}})$; finally, the close-up of $\mathbf{I}$ in the pixel area surrounded by the white box.
• Figure 3: Example of forgery creation process. From left to right: the donor image $\mathbf{I}_D$; the edited version of the donor image $E(\mathbf{I}_D)$; the target image $\mathbf{I}_T$; the locally manipulated target image $\bar{\mathbf{I}}_T$. The splicing is performed by copying a region of the edited donor image into the target one. The manipulation area is surrounded by the red contour.
• Figure 4: Sketch of the tackled counter-forensic task. Given a manipulated amplitude image $\bar{\mathbf{I}}$ and some information on the SAR system used to acquire the original data, we aim at producing an amplitude image $\tilde{\mathbf{I}}$ in which the manipulation traces have been concealed.
• Figure 5: Scheme of the proposed counter-forensic attack. By exploiting information of the original SAR system, we recreate the effect of a SAR acquisition through the same SAR system (or an estimate of it) used to acquire pristine data. To do so, we apply despeckling to the input manipulated image, then we simulate a re-acquisition by injecting speckle noise and filtering with the end-to-end SAR system. Finally, we match the histogram of the produced amplitude image with that of the manipulated amplitude image $\bar{\mathbf{I}}$.