Ada3Diff: Defending against 3D Adversarial Point Clouds via Adaptive Diffusion

TL;DR

Ada3Diff introduces a diffusion-based, distortion-aware defense framework that effectively restores adversarial 3D point clouds, improving robustness against various attack types without heavy computational costs or reliance on specific priors.

Contribution

The paper proposes a novel distortion-aware diffusion model with a tailored intensity estimator for defending against diverse 3D adversarial attacks.

Findings

Effective defense against adaptive attacks with different noise levels.

Restores adversarial point clouds to clean data distribution.

Enhances robustness of 3D recognition models.

Abstract

Deep 3D point cloud models are sensitive to adversarial attacks, which poses threats to safety-critical applications such as autonomous driving. Robust training and defend-by-denoising are typical strategies for defending adversarial perturbations. However, they either induce massive computational overhead or rely heavily upon specified priors, limiting generalized robustness against attacks of all kinds. To remedy it, this paper introduces a novel distortion-aware defense framework that can rebuild the pristine data distribution with a tailored intensity estimator and a diffusion model. To perform distortion-aware forward diffusion, we design a distortion estimation algorithm that is obtained by summing the distance of each point to the best-fitting plane of its local neighboring points, which is based on the observation of the local spatial properties of the adversarial point cloud. By iterative diffusion and reverse denoising, the perturbed point cloud under various distortions can be restored back to a clean distribution. This approach enables effective defense against adaptive attacks with varying noise budgets, enhancing the robustness of existing 3D deep recognition models.