Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
关于 JoVE
概览领导团队博客JoVE 帮助中心
作者
出版流程编辑委员会范围与政策同行评审常见问题投稿
图书馆员
用户评价订阅访问资源图书馆顾问委员会常见问题
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序
Same author

Adaptive diagnostic reasoning framework for pathology with multimodal large language models.

Communications medicine·2026
Same author

Mechanistic Language Modeling and Oxygenated 3D Screening Reveal Berberine and Enzalutamide Synergy in Resistant Prostate Cancer.

bioRxiv : the preprint server for biology·2026
Same author

A refined reweighing technique for nondiscriminatory classification.

PloS one·2024
Same author

Phenotyping senescent mesenchymal stromal cells using AI image translation.

Current research in biotechnology·2023
Same author

Stochastic Optimization for Nonconvex Problem With Inexact Hessian Matrix, Gradient, and Function.

IEEE transactions on neural networks and learning systems·2023
Same author

Trustworthy in silico cell labeling via ensemble-based image translation.

Biophysical reports·2023

相关实验视频

Updated: Jul 11, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

564

时间混合用于防御深度动作识别模型对抗对手攻击.

Jaehui Hwang1, Huan Zhang2, Jun-Ho Choi1

  • 1School of Integrated Technology, Yonsei University, Republic of Korea.

Neural networks : the official journal of the International Neural Network Society
|November 5, 2023
PubMed
概括

动作识别模型通过不太依赖运动来泛化. 时间混合通过破坏视频干扰来防御敌对攻击,为3D CNNs提供了一种新的防御方式,无需重新训练.

关键词:
行动认可 行动认可敌对的攻击/防御.在动作识别中的时间信息.

更多相关视频

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis
05:41

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

9.4K
Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception
05:48

Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception

Published on: August 9, 2024

1.5K

相关实验视频

Last Updated: Jul 11, 2025

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

564
A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis
05:41

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

9.4K
Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception
05:48

Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception

Published on: August 9, 2024

1.5K

科学领域:

  • 计算机视觉 计算机视觉
  • 机器学习 机器学习
  • 深度学习 (Deep Learning) 是一种深度学习.

背景情况:

  • 卷积神经网络 (CNN) 在基于视频的动作识别方面表现出色.
  • 这些模型的概括机制仍然不太了解.
  • 现有的模型可能无法充分利用运动动态.

研究的目的:

  • 研究动作识别模型的概括机制.
  • 识别对敌对攻击的漏洞.
  • 制定一种新的防御策略来应对此类攻击.

主要方法:

  • 分析动作识别模型的稳定性,以框架顺序随机化.
  • 检查运动信息和单调性的作用.
  • 提出一种基于视频的时间混合的防御方法.

主要成果:

  • 动作识别模型显示出意想不到的稳定性来框架顺序随机化.
  • 运动信息的利用比这些模型预期的要少.
  • 敌对扰动对视频中的时间干扰很敏感.

结论:

  • 动作识别模型对运动的依赖程度低于预期,有助于稳定性.
  • 时间混合是对3D CNNs的对抗性攻击的有效防御.
  • 这种防御方法不需要额外的培训,提供了一个实际的解决方案.