IF-Defense: 3D Adversarial Point Cloud Defense via Implicit Function based Restoration

Ziyi Wu, Yueqi Duan, He Wang, Qingnan Fan, Leonidas Guibas

Keywords: Point cloud, adversarial defense, implicit function

Abstract: Point cloud is an important 3D data representation widely used in many essential applications. Leveraging deep neural networks, recent works have shown great success in processing 3D point clouds. However, those deep neural networks are vulnerable to various 3D adversarial attacks, which can be summarized as two primary types: point perturbation that affects local point distribution, and surface distortion that causes dramatic changes in geometry. In this paper, we propose a novel 3D adversarial point cloud defense method leveraging implicit function based restoration (IF-Defense) to address both the aforementioned attacks. It is composed of two steps: 1) it predicts an implicit function that captures the clean shape through a surface recovery module, and 2) restores a clean and complete point cloud via minimizing the difference between the attacked point cloud and the predicted implicit function under geometry- and distribution- aware constraints. Our experimental results show that IF-Defense achieves the state-of-the-art defense performance against all existing adversarial attacks on PointNet, PointNet++, DGCNN and PointConv. Comparing with previous methods, IF-Defense presents 20.02% improvement in classification accuracy against salient point dropping attack and 16.29% against LG-GAN attack on PointNet.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

One-sentence Summary: We propose a 3D adversarial defense method through implicit function based point cloud restoration, which consistently outperforms existing method against various attacks on four network architectures.

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/arxiv:2010.05272/code)

Reviewed Version (pdf): https://openreview.net/references/pdf?id=goqNniT1_