OD-SHIELD: Convolutional Autoencoder-Based Defense Against Adversarial Patch Attacks in Object Detection

In the evolving landscape of deep neural network security, adversarial patch attacks present a serious challenge for object detection systems. We introduce <sc>OD-Shield</sc>, a novel defense approach that employs a convolutional autoencoder framework to detect and remove anomalous regio...

Full description

Saved in:
Bibliographic Details
Main Authors: Byeongchan Kim, Heemin Kim, Minjung Kang, Hyunjee Nam, Sunghwan Park, Jaewoo Lee, Il-Youp Kwak
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11021559/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849332394340909056
author Byeongchan Kim
Heemin Kim
Minjung Kang
Hyunjee Nam
Sunghwan Park
Jaewoo Lee
Il-Youp Kwak
author_facet Byeongchan Kim
Heemin Kim
Minjung Kang
Hyunjee Nam
Sunghwan Park
Jaewoo Lee
Il-Youp Kwak
author_sort Byeongchan Kim
collection DOAJ
description In the evolving landscape of deep neural network security, adversarial patch attacks present a serious challenge for object detection systems. We introduce <sc>OD-Shield</sc>, a novel defense approach that employs a convolutional autoencoder framework to detect and remove anomalous regions in patched images, subsequently reconstructing these regions to mitigate adversarial effects. The reconstructed images are then processed by the object detector, thereby restoring reliable performance under diverse attack scenarios. Distinctly model-agnostic, <sc>OD-Shield</sc> operates as a pre-processing step and can be applied to a wide range of tasks&#x2014;including image classification and object detection&#x2014;without compromising the fidelity of the original image. Experiments on benchmark datasets (COCO, Visdrone, and Argoverse) reveal that <sc>OD-Shield</sc> outperforms existing defenses by 13%&#x2013;47% on COCO, highlighting its effectiveness in addressing critical security vulnerabilities. This work not only tackles the immediate threat of adversarial patches but also lays the foundation for future research into adaptive, resilient defense mechanisms that keep pace with evolving adversarial tactics.
format Article
id doaj-art-a9bf8343286e4310a10e2a6d8e90f774
institution Kabale University
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-a9bf8343286e4310a10e2a6d8e90f7742025-08-20T03:46:12ZengIEEEIEEE Access2169-35362025-01-0113982429825210.1109/ACCESS.2025.357626411021559OD-SHIELD: Convolutional Autoencoder-Based Defense Against Adversarial Patch Attacks in Object DetectionByeongchan Kim0https://orcid.org/0009-0005-4280-4802Heemin Kim1https://orcid.org/0009-0007-1329-2240Minjung Kang2https://orcid.org/0009-0009-3857-1750Hyunjee Nam3https://orcid.org/0009-0004-5334-6416Sunghwan Park4https://orcid.org/0000-0002-0253-110XJaewoo Lee5https://orcid.org/0000-0001-5887-2184Il-Youp Kwak6https://orcid.org/0000-0002-7117-7669Department of Statistics and Data Science, Chung-Ang University, Seoul, South KoreaDepartment of Statistics and Data Science, Chung-Ang University, Seoul, South KoreaDepartment of Statistics and Data Science, Chung-Ang University, Seoul, South KoreaDepartment of Statistics and Data Science, Chung-Ang University, Seoul, South KoreaDepartment of Security Convergence Science, Chung-Ang University, Seoul, South KoreaDepartment of Industrial Security, Chung-Ang University, Seoul, South KoreaDepartment of Statistics and Data Science, Chung-Ang University, Seoul, South KoreaIn the evolving landscape of deep neural network security, adversarial patch attacks present a serious challenge for object detection systems. We introduce <sc>OD-Shield</sc>, a novel defense approach that employs a convolutional autoencoder framework to detect and remove anomalous regions in patched images, subsequently reconstructing these regions to mitigate adversarial effects. The reconstructed images are then processed by the object detector, thereby restoring reliable performance under diverse attack scenarios. Distinctly model-agnostic, <sc>OD-Shield</sc> operates as a pre-processing step and can be applied to a wide range of tasks&#x2014;including image classification and object detection&#x2014;without compromising the fidelity of the original image. Experiments on benchmark datasets (COCO, Visdrone, and Argoverse) reveal that <sc>OD-Shield</sc> outperforms existing defenses by 13%&#x2013;47% on COCO, highlighting its effectiveness in addressing critical security vulnerabilities. This work not only tackles the immediate threat of adversarial patches but also lays the foundation for future research into adaptive, resilient defense mechanisms that keep pace with evolving adversarial tactics.https://ieeexplore.ieee.org/document/11021559/Adversarial patch attackadversarial patch defensepre-processing defenseobject detection
spellingShingle Byeongchan Kim
Heemin Kim
Minjung Kang
Hyunjee Nam
Sunghwan Park
Jaewoo Lee
Il-Youp Kwak
OD-SHIELD: Convolutional Autoencoder-Based Defense Against Adversarial Patch Attacks in Object Detection
IEEE Access
Adversarial patch attack
adversarial patch defense
pre-processing defense
object detection
title OD-SHIELD: Convolutional Autoencoder-Based Defense Against Adversarial Patch Attacks in Object Detection
title_full OD-SHIELD: Convolutional Autoencoder-Based Defense Against Adversarial Patch Attacks in Object Detection
title_fullStr OD-SHIELD: Convolutional Autoencoder-Based Defense Against Adversarial Patch Attacks in Object Detection
title_full_unstemmed OD-SHIELD: Convolutional Autoencoder-Based Defense Against Adversarial Patch Attacks in Object Detection
title_short OD-SHIELD: Convolutional Autoencoder-Based Defense Against Adversarial Patch Attacks in Object Detection
title_sort od shield convolutional autoencoder based defense against adversarial patch attacks in object detection
topic Adversarial patch attack
adversarial patch defense
pre-processing defense
object detection
url https://ieeexplore.ieee.org/document/11021559/
work_keys_str_mv AT byeongchankim odshieldconvolutionalautoencoderbaseddefenseagainstadversarialpatchattacksinobjectdetection
AT heeminkim odshieldconvolutionalautoencoderbaseddefenseagainstadversarialpatchattacksinobjectdetection
AT minjungkang odshieldconvolutionalautoencoderbaseddefenseagainstadversarialpatchattacksinobjectdetection
AT hyunjeenam odshieldconvolutionalautoencoderbaseddefenseagainstadversarialpatchattacksinobjectdetection
AT sunghwanpark odshieldconvolutionalautoencoderbaseddefenseagainstadversarialpatchattacksinobjectdetection
AT jaewoolee odshieldconvolutionalautoencoderbaseddefenseagainstadversarialpatchattacksinobjectdetection
AT ilyoupkwak odshieldconvolutionalautoencoderbaseddefenseagainstadversarialpatchattacksinobjectdetection