Universal Spectral Adversarial Attacks [Paper]

This repository is the PyTorch implementation of the paper:

Universal Spectral Adversarial Attacks for Deformable Shapes
Arianna Rampini, Franco Pestarini, Luca Cosmo, Simone Melzi, Emanuele Rodolà
CVPR, 2021.

To visualize and/or replicate the paper adversarial samples, please take a look at the included Colab Notebook run_adversarial_attack.ipynb.

Data

The shapes used in this example (in datasets) are from the SMAL dataset.