Cargando…
Local imperceptible adversarial attacks against human pose estimation networks
Deep neural networks are vulnerable to attacks from adversarial inputs. Corresponding attack research on human pose estimation (HPE), particularly for body joint detection, has been largely unexplored. Transferring classification-based attack methods to body joint regression tasks is not straightfor...
Autores principales: | , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Springer Nature Singapore
2023
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10661673/ https://www.ncbi.nlm.nih.gov/pubmed/37985638 http://dx.doi.org/10.1186/s42492-023-00148-1 |
_version_ | 1785148474151403520 |
---|---|
author | Liu, Fuchang Zhang, Shen Wang, Hao Yan, Caiping Miao, Yongwei |
author_facet | Liu, Fuchang Zhang, Shen Wang, Hao Yan, Caiping Miao, Yongwei |
author_sort | Liu, Fuchang |
collection | PubMed |
description | Deep neural networks are vulnerable to attacks from adversarial inputs. Corresponding attack research on human pose estimation (HPE), particularly for body joint detection, has been largely unexplored. Transferring classification-based attack methods to body joint regression tasks is not straightforward. Another issue is that the attack effectiveness and imperceptibility contradict each other. To solve these issues, we propose local imperceptible attacks on HPE networks. In particular, we reformulate imperceptible attacks on body joint regression into a constrained maximum allowable attack. Furthermore, we approximate the solution using iterative gradient-based strength refinement and greedy-based pixel selection. Our method crafts effective perceptual adversarial attacks that consider both human perception and attack effectiveness. We conducted a series of imperceptible attacks against state-of-the-art HPE methods, including HigherHRNet, DEKR, and ViTPose. The experimental results demonstrate that the proposed method achieves excellent imperceptibility while maintaining attack effectiveness by significantly reducing the number of perturbed pixels. Approximately 4% of the pixels can achieve sufficient attacks on HPE. |
format | Online Article Text |
id | pubmed-10661673 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2023 |
publisher | Springer Nature Singapore |
record_format | MEDLINE/PubMed |
spelling | pubmed-106616732023-11-21 Local imperceptible adversarial attacks against human pose estimation networks Liu, Fuchang Zhang, Shen Wang, Hao Yan, Caiping Miao, Yongwei Vis Comput Ind Biomed Art Original Article Deep neural networks are vulnerable to attacks from adversarial inputs. Corresponding attack research on human pose estimation (HPE), particularly for body joint detection, has been largely unexplored. Transferring classification-based attack methods to body joint regression tasks is not straightforward. Another issue is that the attack effectiveness and imperceptibility contradict each other. To solve these issues, we propose local imperceptible attacks on HPE networks. In particular, we reformulate imperceptible attacks on body joint regression into a constrained maximum allowable attack. Furthermore, we approximate the solution using iterative gradient-based strength refinement and greedy-based pixel selection. Our method crafts effective perceptual adversarial attacks that consider both human perception and attack effectiveness. We conducted a series of imperceptible attacks against state-of-the-art HPE methods, including HigherHRNet, DEKR, and ViTPose. The experimental results demonstrate that the proposed method achieves excellent imperceptibility while maintaining attack effectiveness by significantly reducing the number of perturbed pixels. Approximately 4% of the pixels can achieve sufficient attacks on HPE. Springer Nature Singapore 2023-11-21 /pmc/articles/PMC10661673/ /pubmed/37985638 http://dx.doi.org/10.1186/s42492-023-00148-1 Text en © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ (https://creativecommons.org/licenses/by/4.0/) . |
spellingShingle | Original Article Liu, Fuchang Zhang, Shen Wang, Hao Yan, Caiping Miao, Yongwei Local imperceptible adversarial attacks against human pose estimation networks |
title | Local imperceptible adversarial attacks against human pose estimation networks |
title_full | Local imperceptible adversarial attacks against human pose estimation networks |
title_fullStr | Local imperceptible adversarial attacks against human pose estimation networks |
title_full_unstemmed | Local imperceptible adversarial attacks against human pose estimation networks |
title_short | Local imperceptible adversarial attacks against human pose estimation networks |
title_sort | local imperceptible adversarial attacks against human pose estimation networks |
topic | Original Article |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10661673/ https://www.ncbi.nlm.nih.gov/pubmed/37985638 http://dx.doi.org/10.1186/s42492-023-00148-1 |
work_keys_str_mv | AT liufuchang localimperceptibleadversarialattacksagainsthumanposeestimationnetworks AT zhangshen localimperceptibleadversarialattacksagainsthumanposeestimationnetworks AT wanghao localimperceptibleadversarialattacksagainsthumanposeestimationnetworks AT yancaiping localimperceptibleadversarialattacksagainsthumanposeestimationnetworks AT miaoyongwei localimperceptibleadversarialattacksagainsthumanposeestimationnetworks |