Transferable and Undefendable Point Cloud Attacks via Medial Axis Transform

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tang, Keke, Gao, Yuze, Peng, Weilong, Wang, Xiaofei, Fang, Meie, Zhu, Peican
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908466440830976
author Tang, Keke
Gao, Yuze
Peng, Weilong
Wang, Xiaofei
Fang, Meie
Zhu, Peican
author_facet Tang, Keke
Gao, Yuze
Peng, Weilong
Wang, Xiaofei
Fang, Meie
Zhu, Peican
contents Studying adversarial attacks on point clouds is essential for evaluating and improving the robustness of 3D deep learning models. However, most existing attack methods are developed under ideal white-box settings and often suffer from limited transferability to unseen models and insufficient robustness against common defense mechanisms. In this paper, we propose MAT-Adv, a novel adversarial attack framework that enhances both transferability and undefendability by explicitly perturbing the medial axis transform (MAT) representations, in order to induce inherent adversarialness in the resulting point clouds. Specifically, we employ an autoencoder to project input point clouds into compact MAT representations that capture the intrinsic geometric structure of point clouds. By perturbing these intrinsic representations, MAT-Adv introduces structural-level adversarial characteristics that remain effective across diverse models and defense strategies. To mitigate overfitting and prevent perturbation collapse, we incorporate a dropout strategy into the optimization of MAT perturbations, further improving transferability and undefendability. Extensive experiments demonstrate that MAT-Adv significantly outperforms existing state-of-the-art methods in both transferability and undefendability. Codes will be made public upon paper acceptance.
format Preprint
id arxiv_https___arxiv_org_abs_2507_18870
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Transferable and Undefendable Point Cloud Attacks via Medial Axis Transform
Tang, Keke
Gao, Yuze
Peng, Weilong
Wang, Xiaofei
Fang, Meie
Zhu, Peican
Computer Vision and Pattern Recognition
Studying adversarial attacks on point clouds is essential for evaluating and improving the robustness of 3D deep learning models. However, most existing attack methods are developed under ideal white-box settings and often suffer from limited transferability to unseen models and insufficient robustness against common defense mechanisms. In this paper, we propose MAT-Adv, a novel adversarial attack framework that enhances both transferability and undefendability by explicitly perturbing the medial axis transform (MAT) representations, in order to induce inherent adversarialness in the resulting point clouds. Specifically, we employ an autoencoder to project input point clouds into compact MAT representations that capture the intrinsic geometric structure of point clouds. By perturbing these intrinsic representations, MAT-Adv introduces structural-level adversarial characteristics that remain effective across diverse models and defense strategies. To mitigate overfitting and prevent perturbation collapse, we incorporate a dropout strategy into the optimization of MAT perturbations, further improving transferability and undefendability. Extensive experiments demonstrate that MAT-Adv significantly outperforms existing state-of-the-art methods in both transferability and undefendability. Codes will be made public upon paper acceptance.
title Transferable and Undefendable Point Cloud Attacks via Medial Axis Transform
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.18870