Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Jul 13, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

Tailoring adversarial attacks on deep neural networks for targeted class manipulation using DeepFool algorithm.

S M Fazle Rabby Labib1, Joyanta Jyoti Mondal2, Meem Arafat Manab3

  • 1School of Data and Sciences, BRAC University, Dhaka, Bangladesh. s.m.fazle.rabby.labib@g.bracu.ac.bd.

Scientific Reports
|March 29, 2025
PubMed
Summary

Related Concept Videos

Survival Tree01:19

Survival Tree

Survival trees are a non-parametric method used in survival analysis to model the relationship between a set of covariates and the time until an event of interest occurs, often referred to as the "time-to-event" or "survival time." This method is particularly useful when dealing with censored data, where the event has not occurred for some individuals by the end of the study period, or when the exact time of the event is unknown.
 Building a Survival Tree
Constructing a survival tree begins...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Comparative performance of AI models and clinicians in evidence-based cardiovascular disease management for people living with HIV: Comparative Study.

Journal of medical Internet research·2026
Same author

HIV infection exacerbates ischemic brain injury through dysregulating phagocytosis and neuroinflammatory signals in macrophages/microglia via C1ql2.

Communications biology·2026
Same author

Rapid <i>de novo</i> assembly of animal-microbe biofilter to mitigate seabed methane leakage.

National science review·2026
Same author

Accumulation of metabolic multimorbidity and its association with mortality: results from a prospective cohort study among people with HIV in China, 2010-2024.

The Lancet regional health. Western Pacific·2026
Same author

Incidence, recurrence, and determinants of sexually transmitted infections among people with HIV: a multicenter cohort study in China, 2010-2024, with implications for integrated HIV-STI prevention across the Western Pacific.

The Lancet regional health. Western Pacific·2026
Same author

Intertidal heterogeneity shapes the metabolomic profile and antialgal potential of Spartina alterniflora.

Marine pollution bulletin·2026

Enhanced Targeted DeepFool (ET DeepFool) improves adversarial attack defenses by allowing targeted misclassifications while preserving image quality. This new method offers greater control over deep neural network vulnerabilities.

Area of Science:

  • Computer Science
  • Artificial Intelligence
  • Machine Learning

Background:

  • Deep neural networks (DNNs) are vulnerable to adversarial attacks, compromising their reliability.
  • Existing methods like DeepFool identify minimal image perturbations for misclassification but lack targeted intervention capabilities.
  • Previous studies often overlooked image quality and confidence thresholds in adversarial attack assessments.

Purpose of the Study:

  • Introduce the Enhanced Targeted DeepFool (ET DeepFool) algorithm for more controlled adversarial attacks.
  • Enable specification of desired misclassification targets and minimum confidence scores.
  • Address limitations in image quality preservation and perturbation minimization.

Main Methods:

  • Developed the Enhanced Targeted DeepFool (ET DeepFool) algorithm, an extension of DeepFool.
Keywords:
Adversarial attackDeep neural networkImage classification

Related Experiment Videos

Last Updated: Jul 13, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

  • Incorporated configurable minimum confidence scores for misclassification.
  • Empirically evaluated ET DeepFool against various DNN architectures, including AlexNet and Vision Transformer.

    • Main Results:

    • ET DeepFool demonstrates superior performance in maintaining image integrity and minimizing perturbations.
    • The algorithm provides enhanced control over the adversarial perturbation process compared to prior methods.
    • Preliminary results indicate varying robustness levels across different DNN models, such as AlexNet and Vision Transformer.

    Conclusions:

    • ET DeepFool offers a more sophisticated approach to understanding and defending against adversarial attacks.
    • The findings highlight the importance of confidence thresholds and image quality in evaluating attack robustness.
    • Discovering differential model robustness has significant implications for secure image recognition systems.