Unleashing the Representational Power of Fourier Shapes for Attacking Infrared Object Detection

2026-05-17

AI Summary

This work addresses the vulnerability of infrared object detection to physical adversarial attacks, where existing shape-based methods face a fundamental trade-off between representational capacity and optimization efficiency, limiting attack performance. To overcome this, the study introduces— for the first time—a learnable Fourier-based shape representation into infrared adversarial attacks. Within an end-to-end differentiable rendering framework, a small set of Fourier coefficients generates pixel-level masks, and optimal adversarial shapes are efficiently discovered by integrating the winding number theorem with gradient-based optimization. This approach resolves the longstanding conflict between expressiveness and optimizability in conventional shape representations. Extensive digital and physical experiments demonstrate significantly enhanced attack effectiveness and robustness: the generated physical patches achieve over 88% success rates at distances beyond 25 meters (at 0.5 confidence) and exhibit strong generalization across varying distances, viewpoints, poses, and target instances.

Abstract

Infrared object detection is crucial for perception in autonomous driving and surveillance but remains vulnerable to physical adversarial attacks. Unlike in the RGB domain, where attacks rely on color texture, infrared attacks must manipulate thermal signatures, making the geometry shape of heat-blocking materials the primary adversarial information carrier. Current shape-based methods suffer from a fundamental trade-off between representational capability and optimization power, limiting their attack effectiveness.In this work, we overcome this dilemma by introducing learnable Fourier shapes to the infrared domain. We utilize an end-to-end differentiable framework where a compact set of Fourier coefficients, defining the shape boundary, is analytically mapped to a pixel-space mask via the winding number theorem. This enables efficient gradient-based optimization to generate potent shapes that cause human targets to evade detection. Extensive digital and physical experiments provide a comprehensive evaluation and validate our superior performance. Our resulting physical patch achieves striking robustness, successfully evading detectors across diverse distances, angles, poses, and individuals, and achieves over 88% attack success rate at distances greater than 25m (conf.=0.5). Code is available at https://github.com/Yongyx99/Fourier-shape-attack.

Problem

• infrared object detection
• physical adversarial attacks
• shape representation
• adversarial robustness
• thermal signatures

Innovation

• Fourier shapes
• infrared object detection
• physical adversarial attack
• differentiable rendering
• winding number theorem