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

SMOC2 mediates GLI1+ mesenchymal cell-driven fibrostenosis and predicts postoperative recurrence in Crohn's disease.

Gut·2026
Same author

SBRT embedded in low-dose RT plus αPD-1 (immuno-EclipseRT, iERT) elicits CD8<sup>+</sup> T cell immunity against bulky tumors via an IFN-I/NK/DC axis.

Nature communications·2026
Same author

Correction: Biotechnological and genetic innovations to enhance sorghum adaptation under climate change.

Frontiers in plant science·2026
Same author

Development of a rapid recombinase polymerase amplification assay for avian polyomaviruses.

Poultry science·2026
Same author

Not All, but the Right Ones: Energy-Guided Representation Learning for Incomplete Multiview Clustering.

IEEE transactions on neural networks and learning systems·2026
Same author

Tumor-Regional Immune Microenvironment: A Critical Factor in the Design of Radiotherapy-Immunotherapy Combination Trials.

International journal of cancer·2026

相关实验视频

Updated: Jul 5, 2025

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

3.8K

基于注意力的3D卷积循环神经网络模型用于多模式情绪识别.

Yiming Du1, Penghai Li1, Longlong Cheng1,2

  • 1School of Integrated Circuit Science and Engineering, Tianjin University of Technology, Tianjin, China.

Frontiers in neuroscience
|January 25, 2024
PubMed
概括

这项研究引入了一种新的3D卷积循环神经网络 (3FACRNN),用于多式情绪识别,通过融合面部和脑电图 (EEG) 数据来提高准确性. 该模型显著提高了情绪识别性能,超过了现有的方法.

关键词:
3D功能构建模块的3D功能构建模块.注意力机制注意力机制卷积神经网络 (CNN) 是一种神经网络.电脑电图 (EEG) 是一种电脑电图.情感识别 情感识别 情感识别多式联运识别多式联运识别

更多相关视频

Exploring the Use of Isolated Expressions and Film Clips to Evaluate Emotion Recognition by People with Traumatic Brain Injury
05:51

Exploring the Use of Isolated Expressions and Film Clips to Evaluate Emotion Recognition by People with Traumatic Brain Injury

Published on: May 15, 2016

9.0K
Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology
09:44

Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology

Published on: March 8, 2024

4.8K

相关实验视频

Last Updated: Jul 5, 2025

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

3.8K
Exploring the Use of Isolated Expressions and Film Clips to Evaluate Emotion Recognition by People with Traumatic Brain Injury
05:51

Exploring the Use of Isolated Expressions and Film Clips to Evaluate Emotion Recognition by People with Traumatic Brain Injury

Published on: May 15, 2016

9.0K
Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology
09:44

Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology

Published on: March 8, 2024

4.8K

科学领域:

  • 人工智能的人工智能
  • 生物医学工程 生物医学工程
  • 认知科学 认知科学

背景情况:

  • 多模式情绪识别对于人机交互和智能医疗保健至关重要.
  • 整合不同的人类模式 (例如视觉,EEG) 进行情绪计算仍然具有挑战性.
  • 现有的方法往往难以有效地融合来自不同模式的信息.

研究的目的:

  • 提出一种新的3D卷积循环神经网络 (3FACRNN) 模型,用于增强多式联络式情绪识别.
  • 使用注意力机制研究视觉和脑电图 (EEG) 数据的融合.
  • 为了提高情绪识别系统的准确性和稳定性.

主要方法:

  • 开发了一个包括视觉网络 (CNN-TCN) 和EEG网络的3FACRNN网络.
  • 在EEG网络中集成频段,空间和时间信息,并构建3D功能.
  • 在卷积循环神经网络 (CRNN) 中使用带注意力和自我注意力模块.
  • 利用多任务损失函数 (Lc) 来调整视觉和EEG模式中的中间特征向量.

主要成果:

  • 3FACRNN模型在DEAP和MAHNOB-HCI数据集上实现了高的识别准确度 (例如,在DEAP上激发的96.75%,值的96.86%).
  • 结果表明,高频马波段 (31-50 Hz) 特别适用于情绪识别.
  • 拟议的方法优于最先进的多式联络式情绪识别方法.

结论:

  • 面部视频和EEG信号的多式融合增强了情绪识别网络的稳定性和准确性.
  • 注意力机制有效地利用来自不同频段和模式的相关信息.
  • 未来的工作将探索稀疏矩阵方法和深度卷积网络,以进一步提高性能.