Adversarially Guided Diffusion for LiDAR Range Image Synthesis

2026-07-08

AI Summary

This work addresses the lack of effective untargeted adversarial attack methods for semantic segmentation of 2D LiDAR range images by introducing diffusion models to this task for the first time. The proposed adversarial-guided diffusion attack integrates segmentation loss into the sampling process to steer the generation of adversarial examples, thereby inducing structured segmentation errors while preserving perceptual realism and data manifold consistency. Experiments on the SemanticKITTI dataset demonstrate that the method not only achieves controllable attack strength and outperforms baseline approaches such as FGSM and SegPGD under white-box settings, but also exhibits strong cross-model transferability.

Abstract

LiDAR semantic segmentation is a key perception task in autonomous driving, where false predictions can affect downstream planning and safety-critical decision-making. Although adversarial attacks, and specifically adversarial examples, have been widely studied for image classification and 3D point cloud segmentation, unrestricted adversarial examples remain largely unexplored in the space of 2D range images, which are projections of 3D point clouds. The proposed method is, to the best of our knowledge, the first diffusion-based unrestricted adversarial attack against 2D range-image segmentation, using adversarial guidance from a segmentation loss. By applying guidance directly during sampling, the method produces unrestricted adversarial examples that remain close to the learned LiDAR data manifold while inducing structured segmentation errors. Experiments on the SemanticKITTI dataset using RangeNet++ and CENet segmentation networks demonstrate that the attack provides adjustable degradation across guidance strengths and transfers across segmentation architectures. Compared with norm-bounded FGSM and SegPGD baselines, the proposed attack offers a distinct effectiveness-realism trade-off, achieving controllable white-box and transfer degradation while maintaining competitive distributional and visual realism.

Problem

• LiDAR range image
• semantic segmentation
• unrestricted adversarial examples
• autonomous driving
• adversarial attack

Innovation

• diffusion-based adversarial attack
• unrestricted adversarial examples
• LiDAR range image
• adversarial guidance
• semantic segmentation