CA-Adv: Curvature-Adaptive Weighted Adversarial 3D Point Cloud Generation Method for Remote Sensing Scenarios
Remote Sensing · 13.03.2026 · 18 (6) · 882
DOI: 10.3390/rs18060882 ↗Kurz gefasst
Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 13.03.2026. Verfasst von Yanwen Sun und 8 weiteren Autor:innen. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of Chin, Fujian Provincial Natural Science Foundation of China, Natural Science Foundation of Xiamen of China.
Abstract (Originalsprache)
Adversarial robustness in 3D point cloud recognition models is a critical concern in remote sensing applications, such as autonomous driving and infrastructure monitoring. Existing adversarial attack methods can compromise model performance; moreover, they often neglect the intrinsic geometric properties of point clouds, leading to perceptually unnatural perturbations that limit their practicality for robustness evaluation in real-world scenarios. To address this, we propose CA-Adv, a novel curvature-adaptive weighted adversarial generation method for 3D point clouds. Our approach first employs Shapley values to assess regional sensitivity and identify salient regions. It then adaptively partitions these regions based on local curvature and assigns perturbation weights accordingly, concentrating the attack on geometrically sensitive areas while preserving overall structural consistency through explicit geometric constraints. Extensive experiments on real-world remote sensing data (KITTI) and synthetic benchmarks (ModelNet40, ShapeNet) demonstrate that CA-Adv achieves a high attack success rate with a minimal perturbation budget. The generated adversarial examples maintain superior visual naturalness and geometric fidelity. The method provides a practical tool for evaluating the robustness of 3D recognition models in applications such as autonomous driving, urban-scale LiDAR perception, and remote sensing point cloud analysis.
Förderung
- National Natural Science Foundation of Chin
- Fujian Provincial Natural Science Foundation of China
- Natural Science Foundation of Xiamen of China
- Jimei University Scientific Research Start-Up Funding Project
- Open Competition for Innovative Projects of Xiamen
Zitieren
APA
Sun, Y., Xiao, S., Liu, W., Huang, M., Cheng, C., Lin, S., Su, J., Wang, Z., & Cai, G. (2026). CA-Adv: Curvature-Adaptive Weighted Adversarial 3D Point Cloud Generation Method for Remote Sensing Scenarios. Remote Sensing, 18(6), 882. https://doi.org/10.3390/rs18060882
Vancouver
Sun Y, Xiao S, Liu W, Huang M, Cheng C, Lin S, et al. CA-Adv: Curvature-Adaptive Weighted Adversarial 3D Point Cloud Generation Method for Remote Sensing Scenarios. Remote Sensing. 2026;18(6):882. doi:10.3390/rs18060882
BibTeX
@article{Sun2026,
author = {Sun, Yanwen and Xiao, Shijia and Liu, Weiquan and Huang, Min and Cheng, Chaozhi and Lin, Shiwei and Su, Jinhe and Wang, Zongyue and Cai, Guorong},
title = {CA-Adv: Curvature-Adaptive Weighted Adversarial 3D Point Cloud Generation Method for Remote Sensing Scenarios},
journal = {Remote Sensing},
year = {2026},
volume = {18},
number = {6},
pages = {882},
doi = {10.3390/rs18060882}
}
EUVIMED