基于HCSP和模糊熵的脑电信号分类
|
摘要
针对共同空间模式特征提取算法(CSP)不能对频域信息进行处理,且在导联数较少的情况下应用效果不佳的问题,提出将希尔伯特-黄变换(HHT)与CSP相结合的算法。在原始脑电信号经过经验模态分解(EMD)后,提取每个导联的前三阶固有模态函数(IMF)及其组合重构信号,利用CSP特征提取,获取2维特征,联合计算信号的自回归模型参数(AR)和模糊熵组成融合特征向量,采用线性判别分类器对提取的特征进行分类。对第二届BCI竞赛提供的数据使用该方法进行特征提取,训练集和测试集分类准确率分别达到90%、88.6%,验证了该算法可有效改善运动想象辨识效果。
The common spatial pattern feature extraction algorithm(CSP)can not deal with the frequency domain information and the poor effects of application exist in the case of a small number of leads,a method combining Hilbert-Huang transform(HHT)with CSP was proposed.The first three intrinsic mode functions(IMF)and the reconstructed signal of each lead were extracted after the empirical mode decomposition(EMD)of the original electroencephalogram(EEG).CSP was used to obtain the two-dimensional feature vector,the autoregressive model parameters(AR)and fuzzy entropy of the EEG were calculated to constitute the feature vectors.The extracted features were classified.The data provided by the second BCI competition and the proposed method for feature extraction were used.Results of experiments show the classification accuracies of 90% and 88.6% are achieved for the training set and test set.The method can affect the motion recognition greatly.
The common spatial pattern feature extraction algorithm(CSP)can not deal with the frequency domain information and the poor effects of application exist in the case of a small number of leads,a method combining Hilbert-Huang transform(HHT)with CSP was proposed.The first three intrinsic mode functions(IMF)and the reconstructed signal of each lead were extracted after the empirical mode decomposition(EMD)of the original electroencephalogram(EEG).CSP was used to obtain the two-dimensional feature vector,the autoregressive model parameters(AR)and fuzzy entropy of the EEG were calculated to constitute the feature vectors.The extracted features were classified.The data provided by the second BCI competition and the proposed method for feature extraction were used.Results of experiments show the classification accuracies of 90% and 88.6% are achieved for the training set and test set.The method can affect the motion recognition greatly.
引文
[1]Long Jinyi,Li Yuanqi,Wang Hongtao,et al.A hybrid brain computer interface to control the direction and speed of a simulated or real wheelchair[J].Neural Systems and Rehabilitation Engineering,2012,20(5):720-729.
[2]ZHAO Limin.Research on EEG feature extraction and classification based on movement imagination[D].Taiyuan:Taiyuan University of Technology,2016(in Chinese).[赵利民.基于运动想象的脑电信号特征提取与分类方法研究[D].太原:太原理工大学,2016.]
[3]Atyabi A,Shic F,Naples A.Mixture of autoregressive modeling orders and its implication on single trial EEG classification[J].Expert Systems with Applications,2016,65:164-180.
[4]Zhou Z,Wan B.Wavelet packet-based independent component analysis for feature extraction from motor imagery EEG of complex movements[J].Clinical Neurophysiology Official Journal of the International Federation of Clinical Neurophysiology,2012,123(9):1779-1788.
[5]ZHENG Shuhua,YAN Chen,WANG Xiangzhou.A repeated bisection CSP feature extraction algorithm of four-class motor imagery EEG[J].Transactions of Beijing Institute of Technology,2016(8):844-850(in Chinese).[郑戍华,闫琛,王向周.一种重复二分CSP4类运动想象脑电信号特征提取算法[J].北京理工大学学报,2016(8):844-850.]
[6]SUN Huiwen,FU Yunfa,XIONG XIN,et al.Identification of EEG induced by motor imagery based on Hilbert-Huang transform[J].Acta Automatica Sinica,2015,41(9):1686-1692(in Chinese).[孙会文,伏云发,熊馨,等.基于HHT运动想象脑电模式识别研究[J].自动化学报,2015,41(9):1686-1692.]
[7]JI Yu.Research of P300processing algorithm based on independent component analysis[D].Hangzhou:Zhejiang University,2013(in Chinese).[计瑜.基于独立分量分析的P300脑电信号处理算法研究[D].杭州:浙江大学,2013.]
[8]YANG Banghua,LU Wenyu,HE Meiyan,et al.Novel feature extraction method for BCI based on WPD and CSP[J].Chinese Journal of Scientific Instrument,2012,33(11):2560-2565(in Chinese).[杨帮华,陆文宇,何美燕,等.脑机接口中基于WPD和CSP的特征提取[J].仪器仪表学报,2012,33(11):2560-2565.]
[9]Park C,Looney D,Rehman NU,et al.Classification of motor imagery BCI using multivariate empirical mode decomposition[J].IEEE Transactions on Neural Systems&Rehabilitation Engineering,2013,21(1):10-22.
[10]LI Conggai.The comparation and application of entropy in the detection of epilepsy[D].Taiyuan:Taiyuan University of Technology,2015(in Chinese).[李聪改.熵在癫痫信号检测中的对比研究与应用[D].太原:太原理工大学,2015.]
[11]TIAN Jing,LUO Zhizeng.Motor imagery EEG feature extraction based on fuzzy entropy[J].Journal of Huazhong University of Science and Technology(Nature Science Edition),2013,41(s1):92-94(in Chinese).[田京,罗志增.基于模糊熵的运动想像脑电信号特征提取[J].华中科技大学学报(自然科学版),2013,41(s1):92-94.]
[12]LI Lijun.Feature extraction and classification of imaginary movements in EEG[D].Guangzhou:South China University of Technology,2012(in Chinese).[李丽君.基于运动想象的脑电信号特征提取及分类算法研究[D].广州:华南理工大学,2012.]
[13]Bajaj V,Pachori RB.Classification of seizure and non-seizure EEG signals using empirical mode decomposition[J].IEEE Transactions on Information Technology in Biomedicine,2012,16(6):1135-1142.
[14]Song X,Yoon SC,Perera V.Adaptive common spatial pattern for single-trial EEG classification in multisubject BCI[C]//International IEEE/EMBS Conference on Neural Engineering.IEEE,2013:411-414.
[15]LIN Haibo,GONG Lu,ZHANG Yi,et al.Feature extraction of EEG signal based on improved HHT and sample entropy[J].Computer Engineering and Design,2015,36(6):1608-1613(in Chinese).[林海波,龚璐,张毅,等.基于改进HHT和样本熵的脑电信号特征提取[J].计算机工程与设计,2015,36(6):1608-1613.]
[2]ZHAO Limin.Research on EEG feature extraction and classification based on movement imagination[D].Taiyuan:Taiyuan University of Technology,2016(in Chinese).[赵利民.基于运动想象的脑电信号特征提取与分类方法研究[D].太原:太原理工大学,2016.]
[3]Atyabi A,Shic F,Naples A.Mixture of autoregressive modeling orders and its implication on single trial EEG classification[J].Expert Systems with Applications,2016,65:164-180.
[4]Zhou Z,Wan B.Wavelet packet-based independent component analysis for feature extraction from motor imagery EEG of complex movements[J].Clinical Neurophysiology Official Journal of the International Federation of Clinical Neurophysiology,2012,123(9):1779-1788.
[5]ZHENG Shuhua,YAN Chen,WANG Xiangzhou.A repeated bisection CSP feature extraction algorithm of four-class motor imagery EEG[J].Transactions of Beijing Institute of Technology,2016(8):844-850(in Chinese).[郑戍华,闫琛,王向周.一种重复二分CSP4类运动想象脑电信号特征提取算法[J].北京理工大学学报,2016(8):844-850.]
[6]SUN Huiwen,FU Yunfa,XIONG XIN,et al.Identification of EEG induced by motor imagery based on Hilbert-Huang transform[J].Acta Automatica Sinica,2015,41(9):1686-1692(in Chinese).[孙会文,伏云发,熊馨,等.基于HHT运动想象脑电模式识别研究[J].自动化学报,2015,41(9):1686-1692.]
[7]JI Yu.Research of P300processing algorithm based on independent component analysis[D].Hangzhou:Zhejiang University,2013(in Chinese).[计瑜.基于独立分量分析的P300脑电信号处理算法研究[D].杭州:浙江大学,2013.]
[8]YANG Banghua,LU Wenyu,HE Meiyan,et al.Novel feature extraction method for BCI based on WPD and CSP[J].Chinese Journal of Scientific Instrument,2012,33(11):2560-2565(in Chinese).[杨帮华,陆文宇,何美燕,等.脑机接口中基于WPD和CSP的特征提取[J].仪器仪表学报,2012,33(11):2560-2565.]
[9]Park C,Looney D,Rehman NU,et al.Classification of motor imagery BCI using multivariate empirical mode decomposition[J].IEEE Transactions on Neural Systems&Rehabilitation Engineering,2013,21(1):10-22.
[10]LI Conggai.The comparation and application of entropy in the detection of epilepsy[D].Taiyuan:Taiyuan University of Technology,2015(in Chinese).[李聪改.熵在癫痫信号检测中的对比研究与应用[D].太原:太原理工大学,2015.]
[11]TIAN Jing,LUO Zhizeng.Motor imagery EEG feature extraction based on fuzzy entropy[J].Journal of Huazhong University of Science and Technology(Nature Science Edition),2013,41(s1):92-94(in Chinese).[田京,罗志增.基于模糊熵的运动想像脑电信号特征提取[J].华中科技大学学报(自然科学版),2013,41(s1):92-94.]
[12]LI Lijun.Feature extraction and classification of imaginary movements in EEG[D].Guangzhou:South China University of Technology,2012(in Chinese).[李丽君.基于运动想象的脑电信号特征提取及分类算法研究[D].广州:华南理工大学,2012.]
[13]Bajaj V,Pachori RB.Classification of seizure and non-seizure EEG signals using empirical mode decomposition[J].IEEE Transactions on Information Technology in Biomedicine,2012,16(6):1135-1142.
[14]Song X,Yoon SC,Perera V.Adaptive common spatial pattern for single-trial EEG classification in multisubject BCI[C]//International IEEE/EMBS Conference on Neural Engineering.IEEE,2013:411-414.
[15]LIN Haibo,GONG Lu,ZHANG Yi,et al.Feature extraction of EEG signal based on improved HHT and sample entropy[J].Computer Engineering and Design,2015,36(6):1608-1613(in Chinese).[林海波,龚璐,张毅,等.基于改进HHT和样本熵的脑电信号特征提取[J].计算机工程与设计,2015,36(6):1608-1613.]