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

Upper Limb Fatigue Information Variation Analysis Based on Parallel Brain and Muscle Functional Networks.

Annals of biomedical engineering·2026
Same author

Network Reconfiguration Underlies Compensatory Muscle Control Across Force Gradients: Parallel Functional Network Evidence from EEG and sEMG.

Annals of biomedical engineering·2026
Same author

Directed network analysis reveals changes in cortical and muscular connectivity caused by different standing balance tasks.

Journal of neural engineering·2022
Same author

Technical Note: Experimental characterization of the dose deposition in parallel MRI-linacs at various magnetic field strengths.

Medical physics·2019
Same author

Comprehensive immune characterization and T-cell receptor repertoire heterogeneity of retroperitoneal liposarcoma.

Cancer science·2019
Same author

Fibroin/peptide co-functionalized calcium titanate nanorods improve osteoinductivity of titanium via mimicking osteogenic niche.

Materials science & engineering. C, Materials for biological applications·2019

相关实验视频

Updated: Jul 19, 2025

Author Spotlight: Using Motor Imagery Brain-Computer Interface to Improve Motor and Cognitive Function in Stroke Patients
09:42

Author Spotlight: Using Motor Imagery Brain-Computer Interface to Improve Motor and Cognitive Function in Stroke Patients

Published on: September 1, 2023

1.3K

EEG-CDILNet:一种轻量级且准确的CNN网络,使用循环扩展卷积来进行运动图像分类.

Tie Liang1,2, Xionghui Yu1,2, Xiaoguang Liu1,2

  • 1Key Laboratory of Digital Medical Engineering of Hebei Province, Hebei University, Baoding 071002, People's Republic of China.

Journal of neural engineering
|August 8, 2023
PubMed
概括

这项研究引入了一种新的轻量级深度学习模型,用于运动图像 (MI) 电脑图像 (EEG) 分类. 拟议的模型实现了高精度,降低了计算成本,平衡了大脑-计算机接口的性能和效率.

关键词:
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.轻量级网络轻量级的网络.运动影像 (MI)

更多相关视频

Author Spotlight: Enhancing Neurorehabilitation Through EEG, Motor Imagery, and Virtual Reality
10:14

Author Spotlight: Enhancing Neurorehabilitation Through EEG, Motor Imagery, and Virtual Reality

Published on: May 10, 2024

1.0K
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
14:08

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

42.7K

相关实验视频

Last Updated: Jul 19, 2025

Author Spotlight: Using Motor Imagery Brain-Computer Interface to Improve Motor and Cognitive Function in Stroke Patients
09:42

Author Spotlight: Using Motor Imagery Brain-Computer Interface to Improve Motor and Cognitive Function in Stroke Patients

Published on: September 1, 2023

1.3K
Author Spotlight: Enhancing Neurorehabilitation Through EEG, Motor Imagery, and Virtual Reality
10:14

Author Spotlight: Enhancing Neurorehabilitation Through EEG, Motor Imagery, and Virtual Reality

Published on: May 10, 2024

1.0K
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
14:08

Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images

Published on: April 13, 2013

42.7K

科学领域:

  • 神经科学是一个神经科学.
  • 机器学习 机器学习
  • 生物医学工程 生物医学工程

背景情况:

  • 使用脑电图 (EEG) 信号进行运动图像 (MI) 分类对于脑电脑接口 (BCI) 至关重要.
  • 深度学习模型提供高精度,但通常需要大量的计算资源,这对实际应用构成挑战.
  • 在MI分类的深度学习中平衡解码性能和计算成本仍然是一个重要的研究障碍.

研究的目的:

  • 开发一种新的,轻量级的端到端卷积神经网络 (CNN),用于准确的运动图像 (MI) 分类.
  • 在基于深度学习的MI分析中解决分类准确性和计算成本之间的权衡问题.
  • 在多个公共EEG数据集上验证拟议模型的有效性.

主要方法:

  • 提出了一个端到端的CNN模型,命名为EEG循环扩展卷积 (CDIL) 网络.
  • 利用深度可分离的卷积来减少参数,并从EEG信号中提取时空特征.
  • 采用循环扩展卷积 (CDIL) 来进行时间变化的深度特征提取和全球平均聚合以减少参数.

主要成果:

  • 获得的平均分类准确率为79.63% (BCIIV2a),94.53% (HGD 4类) 和87.82% (BCIIV2b).
  • 与其他轻量级模型相比,在解码性能和计算成本之间取得了更好的平衡.
  • 废弃实验和特征可视化证实了模型的结构可行性和有效性.

结论:

  • 拟议的CNN模型在显著减少计算资源的情况下提供高MI分类准确性.
  • EEG-CDIL网络为运动图像分类研究中的实际应用提供了可行的解决方案.
  • 这种轻量级的深度学习方法有效地解决了基于EEG的BCI计算成本的挑战.