Search papers, labs, and topics across Lattice.
This paper introduces VFACamou, an end-to-end framework for generating adversarial camouflage that adapts to dynamic physical environments, specifically under UAV reconnaissance. By integrating UV-volume rendering with a diffusion-based texture generator, the method achieves consistent visual performance across varying scales, poses, and lighting conditions, while a novel illumination color consistency estimator ensures natural integration with the background. Extensive experiments show that VFACamou significantly reduces detection rates by maintaining high perceptual naturalness and robust attack performance against multiple detectors.
Achieving stable adversarial camouflage in dynamic environments, VFACamou reduces human detection rates while preserving natural appearance, a breakthrough for physical evasion tactics.
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.
