PCV: A Point Cloud-Based Network Verifier

TL;DR

This paper introduces PCV, a novel verifier for 3D point cloud networks like PointNet, which assesses robustness against adversarial attacks by generating and analyzing perturbed inputs.

Contribution

The paper presents the first point cloud-based network verifier capable of layer-wise robustness verification for 3D classifiers like PointNet.

Findings

PointNet's robustness is compromised by hybrid reverse signed adversarial attacks.

The verifier effectively measures accuracy impact under various perturbations.

Hybrid reverse signed attack significantly affects PointNet's resilience.

Abstract

3D vision with real-time LiDAR-based point cloud data became a vital part of autonomous system research, especially perception and prediction modules use for object classification, segmentation, and detection. Despite their success, point cloud-based network models are vulnerable to multiple adversarial attacks, where the certain factor of changes in the validation set causes significant performance drop in well-trained networks. Most of the existing verifiers work perfectly on 2D convolution. Due to complex architecture, dimension of hyper-parameter, and 3D convolution, no verifiers can perform the basic layer-wise verification. It is difficult to conclude the robustness of a 3D vision model without performing the verification. Because there will be always corner cases and adversarial input that can compromise the model's effectiveness. In this project, we describe a point cloud-based network verifier that successfully deals state of the art 3D classifier PointNet verifies the robustness by generating adversarial inputs. We have used extracted properties from the trained PointNet and changed certain factors for perturbation input. We calculate the impact on model accuracy versus property factor and can test PointNet network's robustness against a small collection of perturbing input states resulting from adversarial attacks like the suggested hybrid reverse signed attack. The experimental results reveal that the resilience property of PointNet is affected by our hybrid reverse signed perturbation strategy