Stealthy and Robust Backdoor Attack against 3D Point Clouds through Additional Point Features

TL;DR

This paper presents SRBA, a stealthy and robust backdoor attack on 3D point clouds that manipulates additional point features, maintaining visual consistency and resisting preprocessing defenses, with high success rates.

Contribution

The paper introduces a novel backdoor attack method that applies feature shifts to 3D point clouds, enhancing stealthiness and robustness against defenses, using Bayesian Optimization for trigger selection.

Findings

Achieves over 94% attack success rate across tests.

Outperforms previous methods under multiple preprocessing defenses.

Maintains visual consistency between poisoned and benign samples.

Abstract

Recently, 3D backdoor attacks have posed a substantial threat to 3D Deep Neural Networks (3D DNNs) designed for 3D point clouds, which are extensively deployed in various security-critical applications. Although the existing 3D backdoor attacks achieved high attack performance, they remain vulnerable to preprocessing-based defenses (e.g., outlier removal and rotation augmentation) and are prone to detection by human inspection. In pursuit of a more challenging-to-defend and stealthy 3D backdoor attack, this paper introduces the Stealthy and Robust Backdoor Attack (SRBA), which ensures robustness and stealthiness through intentional design considerations. The key insight of our attack involves applying a uniform shift to the additional point features of point clouds (e.g., reflection intensity) widely utilized as part of inputs for 3D DNNs as the trigger. Without altering the geometric information of the point clouds, our attack ensures visual consistency between poisoned and benign samples, and demonstrate robustness against preprocessing-based defenses. In addition, to automate our attack, we employ Bayesian Optimization (BO) to identify the suitable trigger. Extensive experiments suggest that SRBA achieves an attack success rate (ASR) exceeding 94% in all cases, and significantly outperforms previous SOTA methods when multiple preprocessing operations are applied during training.