CDUPatch: Color-Driven Universal Adversarial Patch Attack for Dual-Modal Visible-Infrared Detectors
This work addresses a security vulnerability in dual-modal object detection systems, which is incremental as it builds on existing adversarial patch attacks by improving cross-modal effectiveness.
The paper tackles the problem of limited attack effectiveness in dual-modal adversarial patches for visible-infrared object detectors by proposing CDUPatch, a universal cross-modal patch attack that leverages color-driven thermal absorption to generate adversarial infrared textures, resulting in outperforming existing methods on benchmark datasets and demonstrating strong transferability in physical tests.
Adversarial patches are widely used to evaluate the robustness of object detection systems in real-world scenarios. These patches were initially designed to deceive single-modal detectors (e.g., visible or infrared) and have recently been extended to target visible-infrared dual-modal detectors. However, existing dual-modal adversarial patch attacks have limited attack effectiveness across diverse physical scenarios. To address this, we propose CDUPatch, a universal cross-modal patch attack against visible-infrared object detectors across scales, views, and scenarios. Specifically, we observe that color variations lead to different levels of thermal absorption, resulting in temperature differences in infrared imaging. Leveraging this property, we propose an RGB-to-infrared adapter that maps RGB patches to infrared patches, enabling unified optimization of cross-modal patches. By learning an optimal color distribution on the adversarial patch, we can manipulate its thermal response and generate an adversarial infrared texture. Additionally, we introduce a multi-scale clipping strategy and construct a new visible-infrared dataset, MSDrone, which contains aerial vehicle images in varying scales and perspectives. These data augmentation strategies enhance the robustness of our patch in real-world conditions. Experiments on four benchmark datasets (e.g., DroneVehicle, LLVIP, VisDrone, MSDrone) show that our method outperforms existing patch attacks in the digital domain. Extensive physical tests further confirm strong transferability across scales, views, and scenarios.