Adversarial Vulnerability of Temporal Feature Networks for Object Detection
원문 초록
Taking into account information across the temporal domain helps to improve environment perception in autonomous driving. However, it has not been studied so far whether temporally fused neural networks are vulnerable to deliberately generated perturbations, i.e. adversarial attacks, or whether temporal history is an inherent defense against them. In this work, we study whether temporal feature networks for object detection are vulnerable to universal adversarial attacks. We evaluate attacks of two types: imperceptible noise for the whole image and locally-bound adversarial patch. In both cases, perturbations are generated in a white-box manner using PGD. Our experiments confirm, that attacking even a portion of a temporal input suffices to fool the network. We visually assess generated perturbations to gain insights into the functioning of attacks. To enhance the robustness, we apply adversarial training using 5-PGD. Our experiments on KITTI and nuScenes datasets demonstrate, that a model robustified via K-PGD is able to withstand the studied attacks while keeping the mAP-based performance comparable to that of an unattacked model.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
DuET: Dual Incremental Object Detection via Exemplar-Free Task Arithmetic
Vision-based Lifting of 2D Object Detections for Automated Driving
Beyond One Shot, Beyond One Perspective: Cross-View and Long-Horizon Distillation for Better LiDAR Representations
Towards Reliable Detection of Empty Space: Conditional Marked Point Processes for Object Detection
A Real-Time System for Egocentric Hand-Object Interaction Detection in Industrial Domains
ECORE: Energy-Conscious Optimized Routing for Deep Learning Models at the Edge
토론
이 논문과 연결된 포럼 글이 아직 없습니다.