SAGA: Spectral Adversarial Geometric Attack on 3D Meshes

TL;DR

This paper introduces SAGA, a spectral domain-based adversarial attack method on 3D mesh autoencoders, demonstrating how to craft deceptive meshes that alter geometric reconstructions while maintaining visual plausibility.

Contribution

The paper presents a novel spectral adversarial attack framework for 3D meshes, focusing on geometric distortions to deceive autoencoders, which was not addressed in prior semantic-level mesh attack studies.

Findings

Effective spectral perturbations deceive autoencoders

Generated meshes maintain visual credibility

Code is publicly available for reproducibility

Abstract

A triangular mesh is one of the most popular 3D data representations. As such, the deployment of deep neural networks for mesh processing is widely spread and is increasingly attracting more attention. However, neural networks are prone to adversarial attacks, where carefully crafted inputs impair the model's functionality. The need to explore these vulnerabilities is a fundamental factor in the future development of 3D-based applications. Recently, mesh attacks were studied on the semantic level, where classifiers are misled to produce wrong predictions. Nevertheless, mesh surfaces possess complex geometric attributes beyond their semantic meaning, and their analysis often includes the need to encode and reconstruct the geometry of the shape. We propose a novel framework for a geometric adversarial attack on a 3D mesh autoencoder. In this setting, an adversarial input mesh deceives the autoencoder by forcing it to reconstruct a different geometric shape at its output. The malicious input is produced by perturbing a clean shape in the spectral domain. Our method leverages the spectral decomposition of the mesh along with additional mesh-related properties to obtain visually credible results that consider the delicacy of surface distortions. Our code is publicly available at https://github.com/StolikTomer/SAGA.