Attack on Scene Flow using Point Clouds

TL;DR

This paper investigates the robustness of scene flow neural networks against adversarial attacks on point clouds, revealing significant vulnerabilities and demonstrating attack effectiveness with substantial error degradation.

Contribution

Introduces the first adversarial white-box attack method specifically designed for scene flow networks on point clouds, highlighting their susceptibility.

Findings

Adversarial examples cause up to 33.7% error increase on datasets

Attacks on single dimensions or color channels significantly impact accuracy

Scene flow networks are more vulnerable than 2D optical flow networks

Abstract

Deep neural networks have made significant advancements in accurately estimating scene flow using point clouds, which is vital for many applications like video analysis, action recognition, and navigation. The robustness of these techniques, however, remains a concern, particularly in the face of adversarial attacks that have been proven to deceive state-of-the-art deep neural networks in many domains. Surprisingly, the robustness of scene flow networks against such attacks has not been thoroughly investigated. To address this problem, the proposed approach aims to bridge this gap by introducing adversarial white-box attacks specifically tailored for scene flow networks. Experimental results show that the generated adversarial examples obtain up to 33.7 relative degradation in average end-point error on the KITTI and FlyingThings3D datasets. The study also reveals the significant impact that attacks targeting point clouds in only one dimension or color channel have on average end-point error. Analyzing the success and failure of these attacks on the scene flow networks and their 2D optical flow network variants shows a higher vulnerability for the optical flow networks. Code is available at https://github.com/aheldis/Attack-on-Scene-Flow-using-Point-Clouds.git.