高级搜索

基于可调Q因子小波变换的识别左右手运动想象脑电模式研究

陈万忠 王晓旭 张涛

引用本文: 陈万忠, 王晓旭, 张涛. 基于可调Q因子小波变换的识别左右手运动想象脑电模式研究[J]. 电子与信息学报, 2019, 41(3): 530-536. doi: 10.11999/JEIT171191 shu
Citation:  Wanzhong CHEN, Xiaoxu WANG, Tao ZHANG. Research of Discrimination Between Left and Right Hand Motor Imagery EEG Patterns Based on Tunable Q-Factor Wavelet Transform[J]. Journal of Electronics and Information Technology, 2019, 41(3): 530-536. doi: 10.11999/JEIT171191 shu

基于可调Q因子小波变换的识别左右手运动想象脑电模式研究

    作者简介: 陈万忠: 男,1963年生,教授,研究方向为生物信息感知和人机交互;
    王晓旭: 女,1993年生,硕士生,研究方向为信号处理和模式识别;
    张涛: 男,1991年生,博士生,研究方向为信号处理和模式识别
    通讯作者: 陈万忠,chenwz@jlu.edu.cn
  • 基金项目: 中央高校基本科研专项资金(451170301193),吉林省科技发展计划自然基金项目(20150101191JC),吉林省产业技术研发项目(2016C025)

摘要: 针对识别左右手运动想象脑电图信号(EEG)模式精度和互信息不高的问题,该文采用基于可调Q因子小波变换(TQWT)算法来处理脑电信号。首先,利用TQWT对脑电图信号进行分解;随后,提取子频带信号的小波系数能量、自回归模型(AR)系数以及分形维数;最后,利用线性判别分析(LDA)对提取的脑电特征进行识别。采用BCI2003和BCI2005竞赛数据对所提出的算法进行验证,4名受试者的最高识别率分别为88.11%, 89.33%, 77.13%和78.80%,最大互信息分别为0.95, 0.96, 0.43和0.45。实验结果表明,所提算法取得了高分类精度及互信息值,验证了其有效性。

English

    1. [1]

      佘青山, 陈希豪, 高发荣. 基于感兴趣脑区LASSO-Granger因果关系的脑电特征提取算法[J]. 电子与信息学报, 2016, 38(5): 1266–1270. doi: 10.11999/JEIT150851
      SHE Qingshan, CHEN Xihao, and GAO Farong. Feature extraction of electroencephalography based on LASSO-Granger causality between brain region of interest[J]. Journal of Electronics &Information Technology, 2016, 38(5): 1266–1270. doi: 10.11999/JEIT150851

    2. [2]

      BALCONI M and MAZZA G. Brain oscillations and BIS/BAS (behavioral inhibition/activation system) effects on processing masked emotional cues. ERS/ERD and coherence measures of alpha band[J]. International Journal of Psychophysiology, 2009, 74(2): 158–165. doi: 10.1016/j.ijpsycho.2009.08.006

    3. [3]

      吕俊, 谢胜利, 章晋龙. 脑-机接口中基于ERS/ERD的自适应空间滤波算法[J]. 电子与信息学报, 2009, 31(2): 314–318.
      LV Jun, XIE Shengli, and ZHANG Jinlong. Adaptive spatial filter based on ERD/ERS for brain-computer interfaces[J]. Journal of Electronics &Information Technology, 2009, 31(2): 314–318.

    4. [4]

      陈强, 陈勋, 余凤琼. 基于独立向量分析的脑电信号中肌电伪迹的去除方法[J]. 电子与信息学报, 2016, 38(11): 2840–2847. doi: 10.11999/JEIT160209
      CHEN Qiang, CHEN Xun, and YU Fengqiong. Removal of muscle artifact from EEG data based on independent vector analysis[J]. Journal of Electronics &Information Technology, 2016, 38(11): 2840–2847. doi: 10.11999/JEIT160209

    5. [5]

      CHEN Minyou, FANG Yonghui, and ZHENG Xufei. Phase space reconstruction for improving the classification of single trial EEG[J]. Biomedical Signal Processing & Control, 2014, 11(1): 10–16. doi: 10.1016/j.bspc.2014.02.002

    6. [6]

      HAMID M and ZABIHOLLAH S M. Improvement of EEG-based motor imagery classification using ring topology-based particle swarm optimization[J]. Biomedical Signal Processing & Control, 2017, 32: 69–75. doi: 10.1016/j.bspc.2016.10.015

    7. [7]

      PATTNAIK S, DASH M, and SABUT S K. DWT-based feature extraction and classification for motor imaginary EEG signals[C]. International Conference on Systems in Medicine and Biology, Kharagpur, India, 2016: 186–201.

    8. [8]

      徐佳琳, 左国坤. 基于互信息与主成分分析的运动想象脑电特征选择算法[J]. 生物医学工程学杂志, 2016, 33(2): 201–207. doi: 10.7507/1001-5515.20160036
      XU Jialin and ZUO Guokun. Motor imagery electroencephalogram feature selection algorithm based on mutual information and principal component analysis[J]. Journal of Biomedical Engineering, 2016, 33(2): 201–207. doi: 10.7507/1001-5515.20160036

    9. [9]

      罗志增, 周镇定, 周瑛. 双树复小波特征在运动想象脑电识别中的应用[J]. 传感技术学报, 2014, 27(5): 575–580. doi: 10.3969/j.issn.1004-1699.2014.05.001
      LUO Zhizeng, ZHOU Zhending, and ZHOU Ying. The application of DTCWT feature in recognition of motor imagery[J]. Journal of Sensors and Actuators, 2014, 27(5): 575–580. doi: 10.3969/j.issn.1004-1699.2014.05.001

    10. [10]

      周瑛. 虚拟场景下运动想象脑电信号识别研究[D]. [硕士论文], 杭州电子科技大学, 2013.
      ZHOU Ying. The research of motor imagery recognition in virtual reality[D]. [Master dissertation], Hangzhou Dianzi University, 2013.

    11. [11]

      AL-QAZZAZ N K, HAMID B M A S, AHMAD S A, et al. Automatic artifact removal in EEG of normal and demented individuals using ICA-WT during working memory tasks[J]. Sensors, 2017, 17(6): 1–25. doi: 10.3390/s17061326

    12. [12]

      GHORBANIAN P, DEVILBISS D M, VERMA A, et al. Identification of resting and active state EEG features of Alzheimer’s disease using discrete wavelet transform[J]. Annals of Biomedical Engineering, 2013, 41(6): 1243–1257. doi: 10.1007/s10439-013-0795-5

    13. [13]

      HASSAN A R and BHUIYAN M I H. An automated method for sleep staging from EEG signals using normal inverse Gaussian parameters and adaptive boosting[J]. Neurocomputing, 2017, 219: 76–87. doi: 10.1016/j.neucom.2016.09.011

    14. [14]

      BENJAMIN B. BCI Competition II[OL]. http://www.bbci.de/competition/ii/, 2003.

    15. [15]

      BENJAMIN B. BCI Competition III[OL]. http://www.bbci.de/competition/iii/, 2005.

    16. [16]

      VIDAURRE C, SCHLOGL A, CABEZA R, et al. A fully on-line adaptive BCI[J]. IEEE Transactions on Biomedical Engineering, 2006, 53(6): 1214–1219. doi: 10.1109/TBME.2006.873542

    17. [17]

      BAYRAM I and SELESNICK I W. Frequency-domain design of overcomplete rational-dilation wavelet transforms[J]. IEEE Transactions on Signal Processing, 2009, 57(8): 2957–2972. doi: 10.1109/TSP.2009.2020756

    18. [18]

      IVAN S. Tunable Q-factor wavelet transform[OL]. http://eeweb.poly.edu/iselesni/TQWT/index.html, 2016.

    19. [19]

      SELESNICK I W. Wavelet transform with tunable Q-factor[J]. IEEE Transactions on Signal Processing, 2011, 59(8): 3560–3575. doi: 10.1109/TSP.2011.2143711

    20. [20]

      AMIN H U, MALIK A S, AHMAD R F, et al. Feature extraction and classification for EEG signals using wavelet transform and machine learning techniques[J]. Australasian Physical & Engineering Sciences in Medicine, 2015, 38(1): 139–149. doi: 10.1007/s13246-015-0333-x

    21. [21]

      LAWHERN V, HAIRSTON W D, MCDOWELL K, et al. Detection and classification of subject-generated artifacts in EEG signals using autoregressive models[J]. Journal of Neuroscience Methods, 2012, 208(2): 181–189. doi: 10.1016/j.jneumeth.2012.05.017

    22. [22]

      PHOTHISONOTHAI M and NAKAGAWA M. EEG-based classification of motor imagery tasks using fractal dimension and neural network for brain-computer interface[J]. IEICE Transactions on Information and Systems, 2008, 91(1): 44–53. doi: 10.1093/ietisy/e91-d.1.44

    23. [23]

      訾艳阳, 胥永刚, 何正嘉. 离散振动信号分形盒维数的改进算法和应用[J]. 机械科学与技术, 2001(3): 373–376. doi: 10.3321/j.issn:1003-8728.2001.03.021
      ZI Yanyang, XU Yonggang, and HE Zhengjia. Fractal box dimension of discrete vibration signals[J]. Mechanical Science and Technology for Aerospace Engineering, 2001(3): 373–376. doi: 10.3321/j.issn:1003-8728.2001.03.021

    24. [24]

      GUPTA S and SAINI H. EEG features extraction using PCA plus LDA approach based on L1-norm for motor imaginary classification[C]. IEEE International Conference on Computational Intelligence and Computing Research, Coimbatore, India, 2015: 1–5.

    25. [25]

      SCHLOGL A, KEINRATH C, SCHERER R, et al. Information transfer of an EEG-based brain computer interface[C]. International IEEE EMBS Conference on Neural Engineering, Capri Island, Italy, 2003: 641–644.

    26. [26]

      FELE-ZORZ G, KAVSEK G, NOVAK-ANTOLIC Z, et al. A comparison of various linear and non-linear signal processing techniques to separate uterine EMG records of term and pre-term delivery groups[J]. Medical & Biological Engineering & Computing, 2008, 46(9): 911–922. doi: 10.1007/s11517-008-0350-y

    1. [1]

      谭平, 刘利枚, 郭璠, 周开军. Chernoff加权分类器框架在运动想象脑-机接口中的应用. 电子与信息学报, 2019, 41(0): 1-7.

    2. [2]

      王晶, 樊宇, 赵鼎, 杨晨, 王刚, 罗积润. 考虑回旋共振增强效应的平面单栅注波互作用线性分析. 电子与信息学报, 2019, 41(0): 1-6.

    3. [3]

      陈光武, 刘孝博, 王迪, 刘射德. 基于改进小波变换的MEMS陀螺信号去噪算法. 电子与信息学报, 2019, 41(5): 1025-1031.

    4. [4]

      余映, 吴青龙, 邵凯旋, 康迂星, 杨鉴. 超复数域小波变换的显著性检测. 电子与信息学报, 2019, 41(9): 2231-2238.

    5. [5]

      孙怡峰, 吴疆, 黄严严, 汤光明. 一种视频监控中基于航迹的运动小目标检测算法. 电子与信息学报, 2019, 41(0): 1-8.

    6. [6]

      杨若男, 张伟涛, 楼顺天. 基于平行因子分析的SIMO-OFDM系统盲信道与符号联合估计算法. 电子与信息学报, 2019, 41(3): 551-555.

    7. [7]

      王艳, 薛改娜, 李顺波, 惠飞飞. 一类新的周期为2pmq阶二元广义分圆序列的线性复杂度. 电子与信息学报, 2019, 41(9): 2151-2155.

    8. [8]

      刘会杰, 高新海, 郭汝江. 一种低副瓣无混叠的线性调频信号时频分析方法. 电子与信息学报, 2019, 41(0): 1-9.

    9. [9]

      梁春燕, 袁文浩, 李艳玲, 夏斌, 孙文珠. 基于判别邻域嵌入算法的说话人识别. 电子与信息学报, 2019, 41(7): 1774-1778.

    10. [10]

      王玉莹, 张志敏, 李宁, 范怀涛, 赵庆超. 高分宽幅SAR系统下的方位多通道运动目标成像算法研究. 电子与信息学报, 2019, 41(0): 1-6.

    11. [11]

      桑海峰, 陈紫珍. 基于双向门控循环单元的3D人体运动预测. 电子与信息学报, 2019, 41(9): 2256-2263.

    12. [12]

      闫贺, 王珏, 黄佳, 王旭东. 基于二维速度搜索的星载SAR运动目标聚焦算法研究. 电子与信息学报, 2019, 41(6): 1287-1293.

    13. [13]

      张鹏飞, 王树, 李盼. 变换光学中的几个数学问题分析及其在隐身毯设计中的应用. 电子与信息学报, 2019, 41(6): 1336-1343.

    14. [14]

      殷礼胜, 唐圣期, 李胜, 何怡刚. 基于整合移动平均自回归和遗传粒子群优化小波神经网络组合模型的交通流预测. 电子与信息学报, 2019, 41(9): 2273-2279.

    15. [15]

      陈潇翔, 邢孟道, 孙光才, 景国彬. 一种超宽带10 GHz微波光子雷达包络与相位联合运动误差估计方法. 电子与信息学报, 2019, 41(5): 1069-1076.

    16. [16]

      王超, 王岩飞, 王琦, 詹学丽. 基于回波序列最小二乘拟合的高分辨率SAR运动目标速度估计. 电子与信息学报, 2019, 41(5): 1055-1062.

    17. [17]

      廖莎莎, 廖柯, 廖希, 刘力. 高形状因子可编程微波光子滤波器集成芯片. 电子与信息学报, 2019, 41(0): 1-8.

    18. [18]

      李云, 唐英, 刘涵霄. 基于Q-Learning算法的毫微微小区功率控制算法. 电子与信息学报, 2019, 41(0): 1-8.

    19. [19]

      杜小妮, 吕红霞, 王蓉. 一类四重和六重线性码的构造. 电子与信息学报, 2019, 41(0): 1-5.

    20. [20]

      王练, 张贺, 张昭, 张勋杨. 基于自适应随机线性网络编码的优先级调度方案. 电子与信息学报, 2019, 41(8): 1861-1868.

  • 图 1  TQWT分解($J\,$=4)

    图 2  S2受试者3类特征的盒图

    图 3  特征组合后得到的识别率结果

    表 1  不同受试者采用单一特征和组合特征所得平均识别率及最高识别率

    受试者特征组合平均识别率(%)最高识别率(%)
    F181.7486.44
    F280.9585.66
    F367.9373.38
    S1F1+F286.1686.90
    F1+F384.7685.47
    F2+F385.0386.89
    F1+F2+F386.4588.11
    F184.2089.04
    F276.5281.06
    F355.8761.20
    S2F1+F287.8589.30
    F1+F387.6388.59
    F2+F380.2281.33
    F1+F2+F387.9689.33
    F166.0871.46
    F266.3068.93
    F355.1658.92
    S3F1+F275.6176.99
    F1+F371.4073.08
    F2+F371.4972.72
    F1+F2+F374.7077.13
    F173.2477.87
    F269.1074.60
    F352.3658.34
    S4F1+F277.6578.79
    F1+F376.1477.24
    F2+F374.0675.25
    F1+F2+F376.7378.80
    下载: 导出CSV

    表 2  本文方法与文献[5,6]得到的最高识别率

    受试者平均值(%)
    S1S2S3S4
    文献[5]90.7185.5373.1876.9581.59
    文献[6]90.7187.4278.8974.6382.91
    本文方法88.1189.3377.1378.8083.34
    下载: 导出CSV

    表 3  本文方法与BCI2003竞赛前3名获胜者、文献[5,6]方法最大互信息

    特征选择最大互信息
    (bit)
    最小错误识别率
    (%)
    BCI2003_1 st小波特征0.6110.71
    BCI2003_2 ndAR谱能量0.4615.71
    BCI2003_3 rdAAR参数模型0.4517.14
    文献[5]方法相空间特征0.639.29
    文献[6]方法小波特征0.819.29
    本文方法组合特征0.9511.89
    下载: 导出CSV

    表 4  不同受试者TQWT参数设置

    受试者QrJ
    S1132
    S2237
    S3132
    S4233
    下载: 导出CSV

    表 5  本文方法的时耗统计(s)

    TQWT过程能量特征AR系数特征分形维数特征分类总时间
    S10.00100.00120.00160.05590.01740.0771
    S20.00220.00100.00150.05360.01660.0749
    S30.00120.00100.00160.05330.01630.0734
    S40.00140.00150.00180.05470.01710.0765
    下载: 导出CSV
  • 加载中
图(3)表(5)
计量
  • PDF下载量:  20
  • 文章访问数:  303
  • HTML全文浏览量:  131
文章相关
  • 通讯作者:  陈万忠, chenwz@jlu.edu.cn
  • 收稿日期:  2017-12-19
  • 录用日期:  2018-12-06
  • 网络出版日期:  2018-12-21
  • 刊出日期:  2019-03-01
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

/

返回文章