AdvPC: Transferable Adversarial Perturbations on 3D Point Clouds

Abstract

Deep neural networks are vulnerable to adversarial attacks, in which imperceptible perturbations to their input lead to erroneous network predictions. This phenomenon has been extensively studied in the image domain, and has only recently been extended to 3D point clouds. In this work, we present novel data-driven adversarial attacks against 3D point cloud networks. We aim to address the following problems in current 3D point cloud adversarial attacks: they do not transfer well between different networks, and they are easy to defend against via simple statistical methods. To this extent, we develop a new point cloud attack (dubbed AdvPC) that exploits the input data distribution by adding an adversarial loss, after Auto-Encoder reconstruction, to the objective it optimizes. AdvPC leads to perturbations that are resilient against current defenses, while remaining highly transferable compared to state-of-the-art attacks. We test AdvPC using four popular point cloud networks: PointNet, PointNet++ (MSG and SSG), and DGCNN. Our proposed attack increases the attack success rate by up to 40% for those transferred to unseen networks (transferability), while maintaining a high success rate on the attacked network. AdvPC also increases the ability to break defenses by up to 38% as compared to other baselines on the ModelNet40 dataset.

Presented at European conference on computer vision (ECCV), 2020. The code is available at https://github.com/ajhamdi/AdvPC

References in corpus (14)

• Adam: A Method for Stochastic Optimization
• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Explaining and Harnessing Adversarial Examples
• Intriguing properties of neural networks
• PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
• Learning Representations and Generative Models for 3D Point Clouds
• Generating Natural Adversarial Examples
• Adversarial Examples Are Not Bugs, They Are Features
• Adversarial Machine Learning at Scale
• Exploring Spatial Context for 3D Semantic Segmentation of Point Clouds
• Cross-Domain Transferability of Adversarial Perturbations
• Adversarial Objects Against LiDAR-Based Autonomous Driving Systems
• Towards Analyzing Semantic Robustness of Deep Neural Networks
• SADA: Semantic Adversarial Diagnostic Attacks for Autonomous Applications

Cited by in corpus (19)

• Secure and Trustworthy Artificial Intelligence-Extended Reality (AI-XR) for Metaverses
• 3D-VField: Adversarial Augmentation of Point Clouds for Domain Generalization in 3D Object Detection
• IF-Defense: 3D Adversarial Point Cloud Defense via Implicit Function based Restoration
• Exploring Adversarial Robustness of Multi-Sensor Perception Systems in Self Driving
• Towards Analyzing Semantic Robustness of Deep Neural Networks
• A Backdoor Attack against 3D Point Cloud Classifiers
• PointGuard: Provably Robust 3D Point Cloud Classification
• 3D Adversarial Augmentations for Robust Out-of-Domain Predictions
• On Adversarial Robustness of 3D Point Cloud Classification under Adaptive Attacks
• 3D Adversarial Attacks Beyond Point Cloud
• Fooling LiDAR Perception via Adversarial Trajectory Perturbation
• Voint Cloud: Multi-View Point Cloud Representation for 3D Understanding
• Adversarial Robustness of Deep Learning: Theory, Algorithms, and Applications
• Hi-ALPS -- An Experimental Robustness Quantification of Six LiDAR-based Object Detection Systems for Autonomous Driving
• PointBA: Towards Backdoor Attacks in 3D Point Cloud
• Random Walks for Adversarial Meshes
• Generating Unrestricted 3D Adversarial Point Clouds
• The art of defense: letting networks fool the attacker
• Explainability-Aware One Point Attack for Point Cloud Neural Networks