Abstract:Deep learning models for point clouds have shown to be vulnerable to adversarial attacks, which have received increasing attention in various safety-critical applications such as autonomous driving, robotics, and surveillance. Existing 3D attack methods generally employ global distance losses to implicitly constrain the point-wise perturbations for optimization. However, these simple losses are quite difficult to accurately measure and restrict the proper 3D geometry as point clouds are highly structured. Although few recent works try to exploit additional shape-aware surface knowledge to globally constrain the point position, they still fail to preserve the detailed point-to-point geometric dependency in different local regions. To this end, in this paper, we propose a novel Multi-grained Geometry-aware Attack (MGA), which explicitly captures the local topology characteristics in different 3D regions for adversarial constraint. Specifically, we first develop multi-scale spectral local filter banks adapting to different 3D object shapes to explore potential geometric structures in local regions. Considering that objects may contain complex geometries, we then extend each filter bank into multi-layer ones to gradually capture the topology contexts of the same region in a coarse-to-fine manner. Hence, the focused local geometric structures will be highlighted in the coefficients calculated by the filtering process. At last, by restricting these coefficients between benign and adversarial samples, our MGA is able to properly measure and preserve the detailed geometry contexts in the whole 3D object with trivial perturbations. Extensive experiments demonstrate that our attack can achieve superior performance on various 3D classification models, with satisfying adversarial imperceptibility and strong resistance to different defense methods.

Robust Geometry-Dependent Attack for 3D Point Clouds

Improving Adversarial Robustness of 3D Point Cloud Classification Models

Explicitly Perceiving and Preserving the Local Geometric Structures for 3D Point Cloud Attack

Generating 3D Adversarial Point Clouds under the Principle of LiDARs

Adversarial Attack and Defense on Point Sets

Generating 3D Adversarial Point Clouds

Point Cloud Attacks in Graph Spectral Domain: When 3D Geometry Meets Graph Signal Processing

Exploring the Devil in Graph Spectral Domain for 3D Point Cloud Attacks

Domain Adaptation on Point Clouds Via Geometry-Aware Implicits

AdvPC: Transferable Adversarial Perturbations on 3D Point Clouds

Shape Prior Guided Attack: Sparser Perturbations on 3D Point Clouds.

Self-Robust 3D Point Recognition via Gather-Vector Guidance

3D Adversarial Attacks Beyond Point Cloud.

Imperceptible Transfer Attack and Defense on 3D Point Cloud Classification

Adaptive Local Adversarial Attacks on 3D Point Clouds for Augmented Reality

Defense-PointNet: Protecting PointNet Against Adversarial Attacks

PAGN: perturbation adaption generation network for point cloud adversarial defense

PointCA: Evaluating the Robustness of 3D Point Cloud Completion Models Against Adversarial Examples

NormalAttack: Curvature-Aware Shape Deformation along Normals for Imperceptible Point Cloud Attack

Robust Structured Declarative Classifiers for 3D Point Clouds: Defending Adversarial Attacks with Implicit Gradients

On Isometry Robustness Of Deep 3d Point Cloud Models Under Adversarial Attacks