摘要
为了迅速、准确地识别运动想象的脑电信号,提出了一种基于平移窗的运动想象脑电信号活动段提取方法。该方法对运动想象脑电信号Mu/Beta节律的事件相关同步化/去同步化(ERS/ERD)特征进行检测,提取ERS/ERD特征明显的时段;再利用统计量进行特征提取,通过Classify分类器进行信号分类。利用2003年BCI竞赛data setⅢ进行测试,分类准确率达到83.5714%。该方法可以评价受试者的脑活动状态,提高运动想象脑电信号的识别准确率,对脑-机接口实时控制系统的研究有一定的帮助。
In order to quickly and accurately identify electroencephalography(EEG)for imagery movement,in this paper,a method is proposed for extracting active segments of EEG signal for imagery movement based on translating window.This method detects the ERS/ERD characteristics of EEG signal,extracts the time periods with obvious ERS/ERD characteristics,uses the statistics to extract feature,and classifies them with the Classify classifier.Using experiment data setⅢof BCI competition in 2003,the classification accuracy rate reaches 83.5714%.This method can evaluate the brain activity state of subjects,improve the recognition accuracy of motor imagery EEG signals,and be beneficial to the research of brain-computer interface real-time control system.
作者
张莉
王凯瑞
李杨
汪清山
蔡靖
王钢
Zhang Li;Wang Kairui;Li Yang;Wang Qingshan;Cai Jing;Wang Gang(College of Instrumentation&Electrical Engineering,Jilin University,Changchun 130026,China;Beihua University,Jilin 132013,China)
出处
《电子技术应用》
2021年第9期35-38,共4页
Application of Electronic Technique
基金
2019年度吉林大学大学生创新训练计划省级项目(201910183793)
吉林省教育厅“十三五”科学技术项目(JJKH20200964KJ)
吉林省科技计划项目(20190303038SF,20190303043SF)。
关键词
脑电信号识别
活动段检测
事件相关同步化/去同步化
Mu/Beta节律
recognition of electroencephalograph
active segments detection
event-related synchronization/desynchronization
Mu/Beta rhythm
