Hiding Local Manipulations on SAR Images: A Counter-Forensic Attack.

Area of Science:

• Remote Sensing
• Digital Forensics
• Image Analysis

Background:

• Synthetic Aperture Radar (SAR) images are increasingly accessible online.
• This accessibility makes SAR data vulnerable to malicious alterations, like inserting or removing targets.
• Current forensic detectors struggle to identify sophisticated manipulations in SAR amplitude images.

Purpose of the Study:

• To demonstrate a novel counter-forensic attack against SAR image manipulation detection.
• To show that expert manipulation can obscure tampering traces, even from advanced detectors.
• To introduce a generalizable and easy-to-apply attack method.

Main Methods:

• The study proposes a "counter-forensic attack" that simulates re-acquisition of the manipulated scene.
• This attack operates as a black-box method, not targeting specific detectors.
• The approach does not require a training phase or adversarial operations.

Main Results:

• The proposed counter-forensic attack effectively eliminates manipulation traces in SAR amplitude images.
• The attack successfully deceives advanced forensic detectors across diverse scenarios.
• The method proved effective against various manipulation operations.

Conclusions:

• Sophisticated counter-forensic attacks can overcome current SAR image forensic detectors.
• Simulating re-acquisition is a viable strategy for concealing malicious alterations in SAR data.
• Further research is needed to develop robust defenses against such advanced attacks.