[2603.26067] R-PGA: Robust Physical Adversarial Camouflage Generation via Relightable 3D Gaussian Splatting

Computer Science > Computer Vision and Pattern Recognition arXiv:2603.26067 (cs) [Submitted on 27 Mar 2026] Title:R-PGA: Robust Physical Adversarial Camouflage Generation via Relightable 3D Gaussian Splatting Authors:Tianrui Lou, Siyuan Liang, Jiawei Liang, Yuze Gao, Xiaochun Cao View a PDF of the paper titled R-PGA: Robust Physical Adversarial Camouflage Generation via Relightable 3D Gaussian Splatting, by Tianrui Lou and 3 other authors View PDF HTML (experimental) Abstract:Physical adversarial camouflage poses a severe security threat to autonomous driving systems by mapping adversarial textures onto 3D objects. Nevertheless, current methods remain brittle in complex dynamic scenarios, failing to generalize across diverse geometric (e.g., viewing configurations) and radiometric (e.g., dynamic illumination, atmospheric scattering) variations. We attribute this deficiency to two fundamental limitations in simulation and optimization. First, the reliance on coarse, oversimplified simulations (e.g., via CARLA) induces a significant domain gap, confining optimization to a biased feature space. Second, standard strategies targeting average performance result in a rugged loss landscape, leaving the camouflage vulnerable to configuration this http URL bridge these gaps, we propose the Relightable Physical 3D Gaussian Splatting (3DGS) based Attack framework (R-PGA). Technically, to address the simulation fidelity issue, we leverage 3DGS to ensure photo-realistic reconstruction a...