Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published December 2021 | metadata_only
Book Section - Chapter

Adversarially Robust 3D Point Cloud Recognition Using Self-Supervisions


3D point cloud data is increasingly used in safety-critical applications such as autonomous driving. Thus, the robustness of 3D deep learning models against adversarial attacks becomes a major consideration. In this paper, we systematically study the impact of various self-supervised learning proxy tasks on different architectures and threat models for 3D point clouds with adversarial training. Specifically, we study MLP-based (PointNet), convolution-based (DGCNN), and transformer-based (PCT) 3D architectures. Through extensive experimentation, we demonstrate that appropriate applications of self-supervision can significantly enhance the robustness in 3D point cloud recognition, achieving considerable improvements compared to the standard adversarial training baseline. Our analysis reveals that local feature learning is desirable for adversarial robustness in point clouds since it limits the adversarial propagation between the point-level input perturbations and the model's final output. This insight also explains the success of DGCNN and the jigsaw proxy task in achieving stronger 3D adversarial robustness.

Additional Information

We appreciate our area chairs and anonymous reviewers for their insightful comments. We thank Qingzhao Zhang for proofreading our manuscript. Jiachen Sun thanks Zhao Su for her considerate care and help during COVIID-19. This project is partially supported by NSF grants CMMI-2038215 and CNS-1930041.

Additional details

August 20, 2023
August 20, 2023