摘要
对运动想象脑电特征进行准确提取和分类是脑-机接口技术研究的重要问题。针对脑电信号非平稳性和非线性特点,提出了一种将小波包熵(WPE)和支持向量机(SVM)相结合的脑电信号识别方法,利用小波包系数能量分布分析脑电时频特性,结合信息熵分析其不确定性和复杂性,并从单次实验中提取运动想象脑电特征;通过支持向量机对特征信号进行分类,采用了一种核函数参数v和误差惩罚因子c的最佳寻优方法,并用互信息(MI)、信噪比(SNR)、最小错分率(MR)等准则对分类器进行评判。测试结果为:想象左右手运动脑电信号识别精度达到90%,M I为0.65 bit,SNR为1.44。结果表明WPE-SVM识别方法能够准确提取脑电本质特征,具有较强的分类性能和抗干扰能力,为大脑运动意识任务分类提供了有效方法,它可以应用于脑-机接口系统中。
Accurate classification of EEG-based left and right hand motor imagery is an important issue in brain-computer interface.Considering that the electroencephalogram（EEG） signal is nonstationary and nonlinear,a new method for EEG signal recognition based on wavelet packet entropy and support vector machines is proposed in this paper.The time-frequency characteristics and uncertainty of EEG signal are analyzed using wavelet packet entropy,and features of motor imagery EEG are extracted from single trial.Then the feature data are classified using support vector machines（SVM）,meanwhile a finding-minimum-search method is proposed to determine the kernel parameter v and trade-off parameterc.Finally,some evaluation criteria including mutual information（MI）,signal-to-noise ratio（SNR） and misclassification rate（MR） are utilized to analyze the performance of the classifier.Test results achieve that the recognition accuracy is 90%,MR is 0.65 bit and SNR is 1.44.Analysis shows that WPE-SVM method can accurately extract the characteristics of EEG,has strong classification performance and anti-interference capability,and provide an effective approach for classification of mental tasks.The proposed method can be applied in BCI system.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2010年第12期2729-2735,共7页
Chinese Journal of Scientific Instrument
基金
国家基础科学人才培养基金(J0730317)资助项目
关键词
脑-机接口
运动想象
小波包熵
支持向量机
互信息
brain-computer interface
motor imagery
wavelet packet entropy
support vector machine
mutual information
