Adversarial attack against urban scene segmentation for autonomous vehicles

X Xu, J Zhang, Y Li, Y Wang, Y Yang… - IEEE Transactions on …, 2020 - ieeexplore.ieee.org
… related works on both scene segmentation and adversarial attack are first … scene segmentation.
In Section IV, we present the experiments on attacking several latest scene segmentation

On the robustness of semantic segmentation models to adversarial attacks

A Arnab, O Miksik, PHS Torr - Proceedings of the IEEE …, 2018 - openaccess.thecvf.com
… of adversarial attacks on modern semantic segmentation … naturally implement recently
proposed adversarial defenses. Our … Moreover, in the shorter term, we show which segmentation

The vulnerability of semantic segmentation networks to adversarial attacks in autonomous driving: Enhancing extensive environment sensing

A Bar, J Lohdefink, N Kapoor… - IEEE Signal …, 2020 - ieeexplore.ieee.org
… -related tasks, such as semantic segmentation of traffic scenes using nothing but the red-green-…
of CNNs used for semantic segmentation with respect to adversarial attacks, and share …

Advspade: Realistic unrestricted attacks for semantic segmentation

G Shen, C Mao, J Yang, B Ray - arXiv preprint arXiv:1910.02354, 2019 - arxiv.org
… This demonstrates the effectiveness of our unrestricted adversarial attacks for segmentation
Segmentation-based urban traffic scene understanding. In BMVC, volume 1, page 2. …

Segpgd: An effective and efficient adversarial attack for evaluating and boosting segmentation robustness

J Gu, H Zhao, V Tresp, PHS Torr - European Conference on Computer …, 2022 - Springer
… We apply SegPGD as the underlying attack method for segmentation adversarial training.
The adversarial training with our SegPGD achieves state-of-the-art performance on the …

Adversarial attacks for image segmentation on multiple lightweight models

X Kang, B Song, X Du, M Guizani - IEEE Access, 2020 - ieeexplore.ieee.org
… , semantic image segmentation is an essential method of scene understanding that can be
… in the adversarial attack was made. 3. Through experiments, we show that this adversarial

Adversarial attacks on YOLACT instance segmentation

Z Zhang, S Huang, X Liu, B Zhang, D Dong - Computers & Security, 2022 - Elsevier
… focuses on instance segmentation understanding of rail drivable scenes. The dataset …
Our objective is to quantify the pixel-wise accuracy of object labels, similar to semantic scene

Adversarial attacks against lidar semantic segmentation in autonomous driving

Y Zhu, C Miao, F Hajiaghajani, M Huai, L Su… - Proceedings of the 19th …, 2021 - dl.acm.org
segmentation result of the adversarialsegmentation network is misled to perceive the
adversarial point cloud as the reference point cloud (the scene we desired to achieve the attack

Non-local context encoder: Robust biomedical image segmentation against adversarial attacks

X He, S Yang, G Li, H Li, H Chang, Y Yu - Proceedings of the AAAI …, 2019 - aaai.org
Segmentation results from our method on adversarial samples are almost the same as …
of our model against adversarial attacks. Under adversarial attacks, our method produces …

Evaluating the robustness of semantic segmentation for autonomous driving against real-world adversarial patch attacks

F Nesti, G Rossolini, S Nair… - Proceedings of the …, 2022 - openaccess.thecvf.com
… The scene-specific attack was performed on the same set of models as defined earlier. … As
described in Section 3.4, the patch is optimized to be adversarial for a specific urban scene by …