留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于脑电图的新型情绪识别特征提取方法

张克军 韩娜 陈欣然 缪睿

张克军, 韩娜, 陈欣然, 缪睿. 基于脑电图的新型情绪识别特征提取方法[J]. 华南师范大学学报(自然科学版), 2019, 51(5): 6-11. doi: 10.6054/j.jscnun.2019078
引用本文: 张克军, 韩娜, 陈欣然, 缪睿. 基于脑电图的新型情绪识别特征提取方法[J]. 华南师范大学学报(自然科学版), 2019, 51(5): 6-11. doi: 10.6054/j.jscnun.2019078
ZHANG Kejun, HAN Na, CHEN Xinran, MIAO Rui. A New Method for Feature Extraction in Emotion Recognition Based on EEG[J]. Journal of South China normal University (Natural Science Edition), 2019, 51(5): 6-11. doi: 10.6054/j.jscnun.2019078
Citation: ZHANG Kejun, HAN Na, CHEN Xinran, MIAO Rui. A New Method for Feature Extraction in Emotion Recognition Based on EEG[J]. Journal of South China normal University (Natural Science Edition), 2019, 51(5): 6-11. doi: 10.6054/j.jscnun.2019078

基于脑电图的新型情绪识别特征提取方法

doi: 10.6054/j.jscnun.2019078
基金项目: 

国家自然科学基金项目 61901096

详细信息
    通讯作者:

    缪睿,副教授,Email:miaorui.research@gmail.com

  • 中图分类号: TP399

A New Method for Feature Extraction in Emotion Recognition Based on EEG

  • 摘要: 基于脑电图(EEG)信号对情感识别计算进行研究.针对脑电图的特征提取难和模型计算难的问题, 提出了一种从EEG信号中获得可靠区别特征的创新方法.该方法将微分熵与线性判别分析(LDA)相结合,可被应用于情绪EEG信号的特征提取.采用3类情绪EEG数据集进行实验,结果表明该方法能够有效提高EEG分类的性能:与原始数据集的结果相比,平均准确度提高了68%,比单独使用微分熵进行特征提取时的准确度高7%.总执行时间结果证明提出的方法具有较低的时间复杂度.研究结果在3类情感脑电图识别领域具有重要的实用价值,可被应用于实际的工程领域.
  • 图  1  5种方法在原始数据集中的精确度

    Figure  1.  The accuracy of the five methods in the original dataset

    图  2  5种方法在基于LDA的微分熵数据集中的精确度

    Figure  2.  The accuracy of the five methods in the LDA-based differential entropy datasets

    图  3  RF方法在原始数据集中的混淆矩阵图

    Figure  3.  The confusion matrix graph of the RF method in the original dataset

    图  4  SVM方法在基于LDA和SVM的微分熵数据集中的混淆矩阵图

    Figure  4.  The confusion matrix graph of the SVM method in differential entropy dataset based on LDA and SVM

    表  1  不同提取方法对RF、SVM分类的预测性能

    Table  1.   Prediction performance of different extraction methods for RF and SVM classification

    EEG特征提取方法 分类方法 精度/% 准确度/% 召回率/% F1/% Kappa系数
    有接使用原始数据集提取 RF 47.9±3.8 48.7±3.2 48.6±3.5 47.4±4.1 0.227±0.051
    SVM 37.5±2.8 39.2±2.4 37.8±3.1 37.2±2.6 0.069±0.042
    LDA法 RF 53.7±2.8 510±4.0 53.0±2.8 52.3±3.4 0.329±0.039
    SVM 47.5±2.8 49.2±2.4 47.8±3.1 47.2±2.6 0.269±0.042
    微分熵法 RF 70.4±2.6 70.0±2.9 70.4±2.9 69.9±2.9 0.556±0.040
    SVM 77.0±3.0 76.8±3.3 77.0±3.2 76.7±3.2 0.654±0.046
    微分熵与LDA结合法 RF 58.2±2.5 59.0±3.7 59.3±2.8 57.3±3.1 0.378±0.039
    SVM 82.5±3.2 80.1±3.4 80.2±3.1 79.9±3.3 0.698±0.049
    下载: 导出CSV

    表  2  4种不同方法实验的时间复杂度

    Table  2.   The complexity of time in the four experiments with different methods

    实验 方法 伽马频段 总频段
    使用原始数据集预测 RF 4.501±0.386 9.544±0.271
    SVM 17.517±0.980 79.412±2.109
    基本LDA法使用原始数据集预测 RF 2.815±0.036 15.534±0.153
    SVM 2.483±0.024 15.182±0.097
    微分熵数据集预测 RF 2.192±0.023 4.908±0.140
    SVM 3.950±0.213 24.780±0.163
    基于LDA法使用微分熵数据集预测 RF 1.301±0.029 4.819±0.143
    SVM 0.928±0.017 4.366±0.079
    下载: 导出CSV
  • [1] RASHID U, NIAZI I, SIGNAL N, et al. An EEG experimental study evaluating the performance of texas instruments ads1299[J]. Sensors, 2018, 18(11):3721/1-18.
    [2] WOLPAW J R, BIRBAUMER N, HEETDERKS W J, et al. Brain-computer interface technology:a review of the first international meeting[J]. IEEE Transactions on Rehabilitation Engineering, 2000, 8(2):164-173. doi: 10.1109/TRE.2000.847807
    [3] JIN J, SELLERS E W, ZHOU S, et al. A P300 brain-computer interface based on a modification of the mismatch negativity paradigm[J]. International Journal of Neural Systems, 2015, 25(3):1550011/1-13. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=f53f49b5cf1e807c25cbc137d04522cb
    [4] COWIE R, DOUGLAS-COWIE E, TSAPATSOULIS N, et al. Emotion recognition in human-computer interaction[J]. IEEE Signal Processing Magazine, 2001, 18(1):32-80. http://d.old.wanfangdata.com.cn/OAPaper/oai_doaj-articles_706f13007a41d7548233e6f6399e17dd
    [5] ATKINSON J, CAMPOS D. Improving BCI-based emotion recognition by combining EEG feature selection and kernel classifiers[J]. Expert Systems with Applications, 2016, 47:35-41. doi: 10.1016/j.eswa.2015.10.049
    [6] 陈东伟, 易子川, 韩娜, 等.基于脑电图信号在复杂场景下的新型联合算法[J].华南师范大学学报(自然科学版), 2018, 50(6):11-16. http://d.old.wanfangdata.com.cn/Periodical/hnsfdx201806002

    CHEN D W, YI Z C, HAN N, et al. A joint algorithm for electroencephalographic signals at complex scenes[J]. Journal of South China Normal University(Natural Science Edition), 2018, 50(6):11-16. http://d.old.wanfangdata.com.cn/Periodical/hnsfdx201806002
    [7] ZHANG A, YANG B, HUANG L. Feature extraction of EEG signals using power spectral entropy[C]//2008 international conference on Biomedical engineering and informatics. Sanya: IEEE, 2008: 435-439.
    [8] BRUNNER C, BILLINGER M, VIDAURRE C, et al. A comparison of univariate, vector, bilinear autoregressive, and band power features for brain-computer interfaces[J]. Medical & Biological Engineering & Computing, 2011, 49(11):1337-1346. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=9d4f3227ad9c2c2a4819c59367dcb82b
    [9] PETRANTONAKIS P C, HADJILEONTIADIS L J. Emotion recognition from brain signals using hybrid adaptive filtering and higher order crossings analysis[J]. IEEE Transactions on Affective Computing, 2010, 1(2):81-97. doi: 10.1109/T-AFFC.2010.7
    [10] DUAN R N, ZHU J Y, LU B L. Differential entropy feature for EEG-based emotion classification[C]//2013 6th International IEEE/EMBS Conference on Neural Engineering (NER). San Diego: IEEE, 2013: 81-84.
    [11] LEE H, CHOI S. PCAcombine with HMM combine with SVM for EEG pattern classification[C]//Proceedings of the Seventh International Symposium on Signal Processing and Its Applications. Paris: IEEE, 2003.
    [12] 唐建.基于SVM-HMM混合模型的癫痫信号的特征提取与识别[D].重庆: 重庆大学, 2016. http://d.wanfangdata.com.cn/Periodical/D01014827

    TANG J. Feature extraction and recognition of Epilesy signals based on SVM-HMM mixture model[D]. Chongqing: Chongqing University, 2016. http://d.wanfangdata.com.cn/Periodical/D01014827
    [13] ALKAN A, KOKLUKAYA E, SUBASI A. Automatic seizure detection in EEG using logistic regression and artificial neural network[J]. Journal of Neuroscience Methods, 2005, 148(2):167-176. doi: 10.1016/j.jneumeth.2005.04.009
    [14] 朱广明.基于EEG和fNIRS的多模态脑-机接口的特征提取与分类方法研究[D].杭州: 杭州电子科技大学, 2017. http://cdmd.cnki.com.cn/Article/CDMD-10336-1017133568.htm

    ZHU G M. EEG and fNIRS based multimodal BCI: feature extraction and classification[D]. Hangzhou: Hangzhou Dianzi University, 2017. http://cdmd.cnki.com.cn/Article/CDMD-10336-1017133568.htm
    [15] YAZDANI A, EBRAHIMI T, HOFFMANN U. Classification of EEG signals using Dempster Shafer theory and a k-nearest neighbor classifier[C]//2009 4th International IEEE/EMBS Conference on Neural Engineering. Antalya: IEEE, 2009: 327-330.
    [16] FRAIWAN L, LWEESY K, KHASAWNEH N, et al. Automated sleep stage identification system based on time-frequency analysis of a single EEG channel and random forest classifier[J]. Computer Methods and Programs in Biomedicine, 2012, 108(1):10-19. doi: 10.1016/j.cmpb.2011.11.005
    [17] 樊海玮, 史双, 张博敏, 等.基于MLP改进型深度神经网络学习资源推荐算法[J/OL].计算机应用研究, 2020, 37(9).http://www.arocmag.com/article/02-2020-09-006.html.

    FAN H W, SHI S, ZHANG B M, et al. Improved deep neural network of learning resource of recommendation algorithm based on MLP[J/OL]. Application Research of Computers, 2020, 37(9). http://www.arocmag.com/article/02-2020-09-006.html.
    [18] GAUTAMA T, MANDIC D P, van HULLE M M. A differential entropy based method for determining the optimal embedding parameters of a signal[C]//2003 IEEE International Conference on Acoustics, Speech, and Signal Processing.[S.l.]: IEEE, 2003: 6-29.
    [19] KAMBHATLA N, LEEN T K. Dimension reduction by local principal component analysis[J]. Neural Computation, 1997, 9(7):1493-1516. doi: 10.1162/neco.1997.9.7.1493
  • 加载中
图(4) / 表(2)
计量
  • 文章访问数:  2030
  • HTML全文浏览量:  1478
  • PDF下载量:  91
  • 被引次数: 0
出版历程
  • 收稿日期:  2017-07-21
  • 刊出日期:  2019-10-25

目录

    /

    返回文章
    返回