多域特征融合的脑电情绪识别方法

打开文本图片集
本文引用格式:,.多域特征融合的脑电情绪识别方法[J].自动化与信息工程,2025,46(5):38-46.
关键词:脑电情绪识别;多域特征融合;三维卷积神经网络;长短时记忆网络;注意力机制中图分类号:TP391.4 文献标志码:A 文章编号:1674-2605(2025)05-0005-09DOI: 10.12475/aie.20250505 开放获取
EEG Emotion Recognition Method for Multi-domain Feature Fusion
SHAO Wenxin DU Yuxiao (Guangdong University of Technology, School of Automation, Guangzhou 51ooo6, China)
Abstract:Toaddress theissueoflimitedemotional featuresand ineffectiveutilizationof inter-electrode channelinformation inEG-basedemotionrecognitionthisstudyproposesanEEGemotionrecognitiomethodformulti-dmainfeaturefusio.First,a thre-diensioalfatuetctuesostuctedasedoEEGaacterstistegatigditialtropaturs,aeecket energyfeatures,andspatialfeaturesbetweenelectrodechanelstofullpreseeectiveiformationinthefrequencytiead spatial domains.Then,forthe3Dfeaturestructure,ahybridEEGemotionrecognitionmodelnamedCLANNisbuiltbycombining 3Dconvolutialeuraletwork(3DC),ongsormemryetork,andatetiomehasm.Experimetalts onthe DEAP(dataset foremotion analysis using physiological signals)demonstrate thatthe CLANN model achievesa binary classificationaccuracyof 94.50% ,improving the performance of EEG-based emotion recognition.
Keywords:EEGemotionrecognition; mult-domainfeature fusion;3Dconvolutional neuralnetwork;long short-termmemor network;attention mechanism
0 引言
情绪作为人类感知外界事物的一种表达方式,在医疗、教育等领域发挥着重要作用。(剩余11833字)