Authors:
Amon Soares de Souza
1
;
Andreas Meißner
2
and
Michaela Geierhos
1
Affiliations:
1
University of the Bundeswehr Munich, Research Institute CODE, Werner-Heisenberg-Weg 39, 85577 Neubiberg, Germany
;
2
ZITiS, Big Data, Zamdorfer Str. 88, 81677 Munich, Germany
Keyword(s):
Adversarial Optimization, Adversarial Attacks, Image Classification.
Abstract:
Adversarial image processing attacks aim to strike a fine balance between pattern visibility and target model error. This balance ideally results in a sample that maintains high visual fidelity to the original image, but forces the model to output the target of the attack, and is therefore particularly susceptible to transformations by post-processing such as compression. JPEG compression, which is inherently non-differentiable and an integral part of almost every web application, therefore severely limits the set of possible use cases for attacks. Although differentiable JPEG approximations have been proposed, they (1) have not been extended to the stronger and less perceptible optimization-based attacks, and (2) have been insufficiently evaluated. Constrained adversarial optimization allows for a strong combination of success rate and high visual fidelity to the original sample. We present a novel robust attack based on constrained optimization and an adaptive compression search. W
e show that our attack outperforms current robust methods for gradient projection attacks for the same amount of applied perturbation, suggesting a more effective trade-off between perturbation and attack success rate. The code is available here: https://github.com/amonsoes/frcw.
(More)