---
_id: '18218'
abstract:
- lang: eng
  text: Deep neural networks are known to be susceptible to adversarial perturbations
    – small perturbations that alter the output of the network and exist under strict
    norm limitations. While such perturbations are usually discussed as tailored to
    a specific input, a universal perturbation can be constructed to alter the model’s
    output on a set of inputs. Universal perturbations present a more realistic case
    of adversarial attacks, as awareness of the model’s exact input is not required.
    In addition, the universal attack setting raises the subject of generalization
    to unseen data, where given a set of inputs, the universal perturbations aim to
    alter the model’s output on out-of-sample data. In this work, we study physical
    passive patch adversarial attacks on visual odometry-based autonomous navigation
    systems. A visual odometry system aims to infer the relative camera motion between
    two corresponding viewpoints, and is frequently used by vision-based autonomous
    navigation systems to estimate their state. For such navigation systems, a patch
    adversarial perturbation poses a severe security issue, as it can be used to mislead
    a system onto some collision course. To the best of our knowledge, we show for
    the first time that the error margin of a visual odometry model can be significantly
    increased by deploying patch adversarial attacks in the scene. We provide evaluation
    on synthetic closed-loop drone navigation data and demonstrate that a comparable
    vulnerability exists in real data. A reference implementation of the proposed
    method and the reported experiments is provided at https://github.com/patchadversarialattacks/patchadversarialattacks.
alternative_title:
- LNCS
article_processing_charge: No
arxiv: 1
author:
- first_name: Yaniv
  full_name: Nemcovsky, Yaniv
  last_name: Nemcovsky
- first_name: Matan
  full_name: Jacoby, Matan
  last_name: Jacoby
- first_name: Alexander
  full_name: Bronstein, Alexander
  id: 58f3726e-7cba-11ef-ad8b-e6e8cb3904e6
  last_name: Bronstein
  orcid: 0000-0001-9699-8730
- first_name: Chaim
  full_name: Baskin, Chaim
  last_name: Baskin
citation:
  ama: 'Nemcovsky Y, Jacoby M, Bronstein AM, Baskin C. Physical passive patch adversarial
    attacks on visual odometry systems. In: <i>16th Asian Conference on Computer Vision</i>.
    Vol 13847. Springer Nature; 2023:518-534. doi:<a href="https://doi.org/10.1007/978-3-031-26293-7_31">10.1007/978-3-031-26293-7_31</a>'
  apa: 'Nemcovsky, Y., Jacoby, M., Bronstein, A. M., &#38; Baskin, C. (2023). Physical
    passive patch adversarial attacks on visual odometry systems. In <i>16th Asian
    Conference on Computer Vision</i> (Vol. 13847, pp. 518–534). Macao, China: Springer
    Nature. <a href="https://doi.org/10.1007/978-3-031-26293-7_31">https://doi.org/10.1007/978-3-031-26293-7_31</a>'
  chicago: Nemcovsky, Yaniv, Matan Jacoby, Alex M. Bronstein, and Chaim Baskin. “Physical
    Passive Patch Adversarial Attacks on Visual Odometry Systems.” In <i>16th Asian
    Conference on Computer Vision</i>, 13847:518–34. Springer Nature, 2023. <a href="https://doi.org/10.1007/978-3-031-26293-7_31">https://doi.org/10.1007/978-3-031-26293-7_31</a>.
  ieee: Y. Nemcovsky, M. Jacoby, A. M. Bronstein, and C. Baskin, “Physical passive
    patch adversarial attacks on visual odometry systems,” in <i>16th Asian Conference
    on Computer Vision</i>, Macao, China, 2023, vol. 13847, pp. 518–534.
  ista: 'Nemcovsky Y, Jacoby M, Bronstein AM, Baskin C. 2023. Physical passive patch
    adversarial attacks on visual odometry systems. 16th Asian Conference on Computer
    Vision. ACCV: Asian Conference on Computer Vision, LNCS, vol. 13847, 518–534.'
  mla: Nemcovsky, Yaniv, et al. “Physical Passive Patch Adversarial Attacks on Visual
    Odometry Systems.” <i>16th Asian Conference on Computer Vision</i>, vol. 13847,
    Springer Nature, 2023, pp. 518–34, doi:<a href="https://doi.org/10.1007/978-3-031-26293-7_31">10.1007/978-3-031-26293-7_31</a>.
  short: Y. Nemcovsky, M. Jacoby, A.M. Bronstein, C. Baskin, in:, 16th Asian Conference
    on Computer Vision, Springer Nature, 2023, pp. 518–534.
conference:
  end_date: 2022-12-08
  location: Macao, China
  name: 'ACCV: Asian Conference on Computer Vision'
  start_date: 2022-12-04
date_created: 2024-10-08T12:51:14Z
date_published: 2023-03-11T00:00:00Z
date_updated: 2024-10-09T12:13:36Z
day: '11'
doi: 10.1007/978-3-031-26293-7_31
extern: '1'
external_id:
  arxiv:
  - '2207.05729'
intvolume: '     13847'
language:
- iso: eng
main_file_link:
- open_access: '1'
  url: https://doi.org/10.48550/arXiv.2207.05729
month: '03'
oa: 1
oa_version: Preprint
page: 518-534
publication: 16th Asian Conference on Computer Vision
publication_identifier:
  eisbn:
  - '9783031262937'
  eissn:
  - 1611-3349
  isbn:
  - '9783031262920'
  issn:
  - 0302-9743
publication_status: published
publisher: Springer Nature
quality_controlled: '1'
related_material:
  link:
  - relation: software
    url: https://github.com/patchadversarialattacks/patchadversarialattacks
scopus_import: '1'
status: public
title: Physical passive patch adversarial attacks on visual odometry systems
type: conference
user_id: 2DF688A6-F248-11E8-B48F-1D18A9856A87
volume: 13847
year: '2023'
...