VFACamou: View-Fused Adversarial Camouflage for Environment-Adaptive Physical Evasion

2026-06-17

AI Summary

This work addresses the challenge of adversarial camouflage failure and unnatural texture appearance in drone-based reconnaissance caused by drastic variations in viewpoint, pose, and illumination. To overcome these limitations, the authors propose an end-to-end wearable adversarial camouflage generation framework that integrates UV volumetric rendering with diffusion models for texture synthesis. The approach incorporates a lighting color consistency estimator, a natural texture loss, and a multi-scale dynamic training strategy to achieve robustness across viewpoints and adaptability to diverse environments. Experimental results demonstrate that the generated camouflages significantly reduce target detectability across multiple state-of-the-art detectors while preserving high visual naturalness and physical attack stability, without exhibiting noticeable artifacts or artificial traces.

Abstract

Adversarial camouflage in the physical world remains highly challenging, particularly under UAV reconnaissance where targets undergo continuous geometric changes and extreme illumination variations. Existing methods either optimize 2D digital perturbations that fail to generalize to dynamic viewpoints or produce visually unnatural textures that cannot be deployed in real scenarios. Therefore, we propose an end-to-end framework for adversarial camouflage generation that automatically produces wearable adversarial patterns and maintains stable attack performance in real physical environments with changing viewpoints, poses, and lighting conditions. Our method integrates UV-volume rendering with a diffusion-based texture generator, enabling consistent appearance under varying scales, poses, and lighting conditions. To ensure environmental realism, we propose an illumination color consistency estimator that extracts dominant background attributes and guides a natural texture loss to align the generated UV texture with the surrounding environment. A multi-scale dynamic training strategy further enhances robustness against viewpoint shifts and body deformation. Extensive experiments across multiple mainstream detectors demonstrate that our method achieves strong and stable physical attack performance while maintaining high perceptual naturalness, reducing human detection rates without introducing unnatural artifacts.

Problem

• adversarial camouflage
• physical evasion
• UAV reconnaissance
• viewpoint variation
• illumination variation

Innovation

• adversarial camouflage
• UV-volume rendering
• diffusion-based texture generation
• illumination consistency
• multi-scale dynamic training