基于三维特征图和改进DenseNet的脑电情绪识别方法  

An EEG emotion recognition method based on 3D feature and improved DenseNet

作　　者：苏靖然 李秋生[1,2] SU Jingran;LI Qiusheng(School of Physics and Electronic Information,Gannan Normal University,Ganzhou,Jiangxi 341000,China;Center for Intelligent Control Engineering Research,Gannan Normal University,Ganzhou,Jiangxi 341000,China)

机构地区：[1]赣南师范大学物理与电子信息学院,江西赣州341000 [2]赣南师范大学智能控制工程技术研究中心,江西赣州341000

出　　处：《石河子大学学报（自然科学版）》2023年第3期381-389,共9页Journal of Shihezi University(Natural Science)

基　　金：国家自然科学基金项目(61561004);江西省教育厅科学技术研究项目(GJJ201408)。

摘　　要：情感作为人脑的高级功能,对人的心理健康状态有很大的影响。为了充分考虑脑电信号的空间信息以及时频信息,更好地实现人机交互,论文提出了1种基于三维特征图的改进DenseNet情绪识别模型。通过提取脑电信号θ、α、β和γ 4个频段的微分熵特征,结合脑电通道电极的位置映射关系,构造三维特征图,最后使用改进DenseNet网络进行二次特征提取与分类。为了验证该方法的有效性,在SEED数据集上进行了包含积极、中性、消极3种情绪的分类实验,单被试者实验和所有被试者实验获得的分类准确率分别达98.51%和98.68%。实验结果表明,三维特征图结合特征重用方法能够得到高精度的分类结果,为情绪识别提供了可以尝试的新方向。Emotion,as a high-level function of the human brain,greatly influences human mental health.To fully consider EEG signals′ spatial information and time-frequency information and realize human-computer interaction better.An improved DenseNet emotion recognition model was proposed based on a three-dimensional feature map.By extracting the differential entropy features of theta,alpha,beta,and gamma frequency bands of EEG signal and combining with the position mapping of EEG channel electrodes,a 3D feature map was constructed.Finally,the improved DenseNet network was used for secondary feature extraction and classification.In order to verify the effectiveness of this method,positive,neutral,and negative classification experiments were carried out on the SEED data set.The classification accuracy of the single-subject experiment and all subjects experiment was 98.51% and 98.68%,respectively.Experimental results showed that 3D feature maps combined with the feature reuse method could obtain high-precision classification results.It provides a compelling new direction for emotion recognition.

关 键 词：脑电信号 电极映射 三维特征图 特征重用 多尺度卷积核 

分 类 号：R318[医药卫生—生物医学工程] TN911.7[医药卫生—基础医学]