Benchmarking the Robustness of Segmentation Methods Against Adversarial Attacks in Breast Ultrasound Segmentation

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This study presents a comprehensive robustness analysis of five segmentation models—UResNet, DeepLabV3, TransUnet, SAM, and Efficient SAM (ESAM)—for breast ultrasound (BUS) image segmentation under adversarial attacks. Each model is initially trained on a clean BUS dataset, followed by systematic evaluation against five widely-used adversarial techniques: FGSM, BIM, PGD, PGDL2, and Jitter. Model performance is quantitatively evaluated using tumor IoU, background IoU, and mean IoU metrics on both clean and adversarial data. Experimental results show that CNN-based models, such as UResNet and DeepLabV3, were more resilient to adversarial perturbations, maintaining higher accuracy compared to transformer-based models like TransUnet, SAM, and the lightweight ESAM, which exhibited significant vulnerability. These findings emphasize the importance of robustness evaluations in medical imaging and other high-stakes applications, where performance degradation can result in serious consequences. This study highlights the need for developing more robust models and effective defense strategies to enhance the reliability of medical image segmentation systems in clinical applications.

Original languageEnglish
Title of host publicationComputational Science and Computational Intelligence - 11th International Conference, CSCI 2024, Proceedings
EditorsHamid R. Arabnia, Leonidas Deligiannidis, Farzan Shenavarmasouleh, Soheyla Amirian, Farid Ghareh Mohammadi
PublisherSpringer Science and Business Media Deutschland GmbH
Pages188-201
Number of pages14
ISBN (Print)9783031949616
DOIs
StatePublished - 2025
Event11th International Conference on Computational Science and Computational Intelligence, CSCI 2024 - Las Vegas, United States
Duration: 11 Dec 202413 Dec 2024

Publication series

NameCommunications in Computer and Information Science
Volume2511 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference11th International Conference on Computational Science and Computational Intelligence, CSCI 2024
Country/TerritoryUnited States
CityLas Vegas
Period11/12/2413/12/24

Keywords

  • Adversarial Attacks
  • Deep Learning
  • Medical Image Segmentation
  • Robustness Analysis

Fingerprint

Dive into the research topics of 'Benchmarking the Robustness of Segmentation Methods Against Adversarial Attacks in Breast Ultrasound Segmentation'. Together they form a unique fingerprint.

Cite this