基于PCA融合PSO-SVM的运动想象脑电信号分类方法

运动想象脑电信号的分类识别是当前脑机接口(BCI)技术面临的难点。针对该问题,提出一种融合主成分分析(PCA)和粒子群优化-支撑向量机(PSO-SVM)的运动想象脑电信号分类方法。首先利用PCA对采集到的高维脑电信号进行分析,剔除其中噪声分量并提取三维反应不同脑电信号差异特性的特征向量。然后利用SVM对特征向量进行分类,同时针对SVM分类性能受核参数影响较大的问题,利用PSO算法的全局寻优能力对其进行优化,从而提升SVM的分类性能。最后采用BCI竞赛中所用Graz数据进行实验,结果表明所提的PCA融合PSO-SVM方法可以获得95.3%的分类性能,在低信噪比条件下具有鲁棒性和较高的应用前景。     

基于LMD和CSP的多域融合脑电信号分类方法

运动想象脑电信号非平稳、非线性和微弱性特征明显,采用传统单一维度特征进行分类时存在识别率低、鲁棒性差的问题。提出一种基于局部均值分解(Local Mean Decomposition, LMD)和共空间模式(Common Spatial Pattern, CSP)的多域融合脑电信号分类方法,采用LMD对运动脑电信号进行自适应分解得到多个乘积分量(Product Function, PF),进而从PF中提取反映不同信号差异特性的12维时-频域特征,将PF作为CSP的多通道数据进行分解,并提取18维空域特征。利用相关向量机(Relevance Vector Machine, RVM)分类器对30维时-频-空域特征进行特征选择和分类识别,在自动确定最优分类特征的同时获得理想的分类结果。基于BCI竞赛数据开展实验,结果表明,所提方法可以获得优于95%的正确分类性能,并且在低信噪比条件下具有较强的噪声稳健性。     

基于分数阶傅里叶变换的运动脑电信号分类方法

运动想象脑电信号作为一种典型的非线性、非平稳信号,在传统基于单一特征提取的分类方法中难以取得理想的分类性能。针对该问题,将分数阶傅里叶变换(Fractional Fourier Transform, FrFT)引入到脑电信号特征提取过程中。首先利用FrFT对信号进行分析,在扩展特征域的同时从不同维度提取信号中的有用信息并构成特征向量,然后利用支持向量机(Support Vector Machine, SVM)分类器对所提取的特征向量进行分类,最后采用Graz数据开展实验。实验结果表明所提方法能够获得高达92.57%的正确分类结果,明显高于传统采用单一特征提取的分类方法。     

基于经验模态分解和SVM的脑电信号分类方法

脑电信号的非线性、非平稳性造成对运动想象脑电信号的分类识别存在特征提取困难、可区分性低以及分类识别性能差等问题。本文提出一种基于经验模态分解(Empirical Mode Decomposition, EMD)和支撑向量机(Support Vector Machine, SVM)的运动想象脑电信号分类方法,充分利用EMD算法在处理非线性、非平稳信号的自适应性以及SVM在小样本条件的高识别性能和强泛化能力。首先利用EMD算法将C3、C4导联信号分解为一系列本征模函数(Intrinsic Mode Function, IMF),然后从IMF的信息和能量等维度提取特征将脑电信号转换至区分性更强的特征域,最后利用SVM进行分类识别。采用国际BCI竞赛2003中的Graz数据进行验证,所提方法可以得到94.6%的正确识别率,为在线脑-机接口系统的研究提供了新的思路。     

基于奇异谱熵的脑电意识任务识别方法的研究

奇异谱分析是脑电信号研究的一种新方法。脑电信号的奇异谱熵可以反映脑电的特征,它有助于研究大脑的动力学行为。时变脑电信号所对应的奇异谱熵时间序列能很好地反映出事件相关去同步(ERD)和事件相关同步(ERS)现象,因此可以提取人脑想象左右手运动任务时的特征,最终利用K-近邻模式分类方法对想象左右手运动任务进行有效的分类决策。最后对国际脑机接口竞赛2003相关数据进行了测试,最高准确率达到85.16%,最大互信息达到0.48。测试结果说明,基于奇异谱熵的脑电信号特征,可以作为脑电意识任务的有效分类依据。     

基于能量特征的脑电信号特征提取与分类

为了快速、有效地提取脑电特征,提高分类正确率,采用带通滤波和小波包分析的方法提取Mu、Beta节律对应的脑电信号,在时域范围内,将信号幅度的平方作为能量特征值;在频域范围内,采用AR模型功率谱估计法所得的功率谱密度作为能量特征值.根据运动想象脑电信号特点,构造左右通道信号能量差值的符号特性作为分类判别依据,进行分类测试,方法简单.初步实验结果表明,所利用的两种方法的分类正确率达87.857%.     

一种新的BP学习算法及在运动想象脑电信号分类识别中的应用

对运动想象(MI)脑电信号的正确分类是决定基于运动想象脑电的脑－机接口(BCI)性能的关键因素。为有效地提取MI脑电信号特征、提高分类正确率,提出一种基于单形进化的BP神经网络优化算法(BPSSSE)并运用于MI脑电信号的识别,提取自相关(AR)模型参数和希尔伯特边际谱作为特征输入,通过单形进化算法优化BP神经网络学习性能,实现对MI脑电信号的分类。测试实验中,对BCI竞赛数据进行左右手分类。结果表明在4s~ 8s时间段内平均分类正确率为80.17%,最高分类正确率为87.14%,证明了本文算法在基于MI脑电的脑机交互控制系统中应用研究的有效性和可行性。     

基于希尔伯特变换联合卷积神经网络的脑电信号识别方法

传统运动想象脑电信号识别方法需要人为提取大量特征,识别性能受研究人员经验影响较大,主观性强;提出一种基于希尔伯特变换(HT)联合卷积神经网络(CNN)的运动想象脑电信号自动识别方法,首先利用HT对原始EEG信号进行分析,实现一维数据向二维幅-相图像转换的同时增加信息提取维度;然后将其作为输入利用CNN层次化的对幅-相二维图像进行理解和解译,自动提取特征并完成分类识别,基于BCI竞赛中所用Graz数据集开展试验,结果表明相对于传统特征提取方法,文章所提算法在低、中、高信噪比条件下均能获得更好的识别性能,具有更强的噪声鲁棒性.     

基于小波变换和AdaBoost极限学习机的癫痫脑电信号分类

针对单一极限学习机(ELM)在癫痫脑电信号研究中分类结果不稳定、泛化能力差的缺陷,提出一种基于互信息(MI)的AdaBoost极限学习机分类算法。该算法将AdaBoost引入到极限学习机中,并嵌入互信息输入变量选择,以强学习器最终的性能作为评价指标,实现对输入变量以及网络模型的优化。利用小波变换(WT)提取脑电信号特征,并结合提出的分类算法对UCI脑电数据集以及波恩大学癫痫脑电数据进行分类。实验结果表明,所提方法相比传统方法以及其他同类型研究,在分类精度和稳定性上有着明显提高,并具有较好的泛化性能。     

基于Normal-Gamma共轭先验优化因子分析的图像去噪算法

采用传统因子分析(Fact Analysis, FA)模型进行噪声抑制时存在因子个数难以确定、噪声抑制后图像质量下降等问题，基于贝叶斯决策理论提出一种Normal-Gamma共轭先验优化FA模型的图像去噪算法，利用Normal-Gamma分布对FA加载因子和隐变量的概率分布建模，一方面对不适定噪声抑制问题正则化，另一方面提高模型的稀疏性，增加参数估计稳定性，采用变分贝叶斯期望最大(Variational Bayesian Expectation Maximum, VBEM)算法对模型求解，自动确定因子个数的同时提升噪声抑制性能。基于标准图像数据集的实验结果表明，所提算法在实现噪声抑制的同时较好地保留了图像的边缘和纹理等细节信息，并且能够明显提升低信噪比条件下的图像识别性能。     

结合非线性全局特征和谱特征的脑电情感识别

针对现有表征情感信息的脑电信号的非线性特征提取不完善的问题,将相空间重构技术引入情感脑电的识别中,提取了在相空间重构下基于轨迹的描述轮廓的三种非线性几何特征作为新的情感脑电特征。结合脑电信号的功率谱熵以及非线性属性特征（近似熵、最大Lyapunov指数、Hurst指数）,提出了基于主成分分析（PCA）的非线性全局特征（非线性几何特征+非线性属性特征）和功率谱熵的融合算法,以支持向量机（SVM）为分类器进行情感识别。结果显示,非线性全局特征能更有效地实现情感识别,二分类情感识别率约90%左右。基于PCA的融合情感特征相比单一特征能达到更佳的情感识别性能,四分类实验中平均识别率可达86.42%。结果表明,非线性全局特征相比非线性属性特征情感识别率有所提高,非线性全局特征以及功率谱熵的结合可以构造出更佳的情感脑电特征参数。     

基于皮尔逊最优电极选择的ADHD患者脑电特征提取及分类研究

事件相关电位(ERP)可用于注意缺陷多动障碍儿童(ADHD)和正常儿童的脑电特征 提取与分类。首先,采用赌博任务范式,采集2 类儿童的脑电信号;其次,基于皮尔逊相关系 数算法选择最优电极,并预处理最优电极脑电信号;然后,提取预处理脑电信号的时域特征(均 值、方差、峰值)和频域特征(Theta 波段功率、Alpha 波段功率);最后,利用传统分类方法支持 向量机(SVM)、自适应增强(AdaBoost)、自举汇聚法(Bagging)、线性判别式分析(LDA)、反向传 播(BP)和组合分类器的分类方法(LDA-SVM,BP-SVM)完成对2 种脑电信号的分类。研究结果 表明,传统方法BP 分类器的分类准确率可达80.52%,组合分类器BP-SVM 的分类准确率可达 88.88%。组合分类方法能提高ADHD 儿童的分类准确率,为基于脑机接口技术的ADHD 神经 反馈康复治疗提供技术支持。     

Measuring saliency of features representing EEG signals using signal-to-noise ratios

Various methodologies of automated diagnosis have been adopted, however the entire process can generally be subdivided into a number of disjoint processing modules: pre-processing, feature extraction/selection, and classification. Features are used to represent patterns with minimal loss of important information. The feature vector, which is comprised of the set of all features used to describe a pattern, is a reduced-dimensional representation of that pattern. Medical diagnostic accuracies can be improved when the pattern is simplified through representation by important features. By identifying a set of salient features, the noise in a classification model can be reduced, resulting in more accurate classification. In this study, a signal-to-noise ratio (SNR) saliency measure was employed to determine saliency of input features of probabilistic neural networks (PNNs) used in classification of electroencephalogram (EEG) signals. In order to extract features representing the EEG signals, eigenvector methods were used. The PNNs used in the EEG signals classification were trained for the SNR screening method. The application results of the SNR screening method to the EEG signals demonstrated that classification accuracies of the PNNs with salient input features are higher than that of the PNNs with salient and non-salient input features.     

Robust Feature Extraction for Continuous Speech Recognition Using the MVDR Spectrum Estimation Method

This paper describes a robust feature extraction technique for continuous speech recognition. Central to the technique is the minimum variance distortionless response (MVDR) method of spectrum estimation. We consider incorporating perceptual information in two ways: 1) after the MVDR power spectrum is computed and 2) directly during the MVDR spectrum estimation. We show that incorporating perceptual information directly into the spectrum estimation improves both robustness and computational efficiency significantly. We analyze the class separability and speaker variability properties of the features using a Fisher linear discriminant measure and show that these features provide better class separability and better suppression of speaker-dependent information than the widely used mel frequency cepstral coefficient (MFCC) features. We evaluate the technique on four different tasks: an in-car speech recognition task, the Aurora-2 matched task, the Wall Street Journal (WSJ) task, and the Switchboard task. The new feature extraction technique gives lower word-error-rates than the MFCC and perceptual linear prediction (PLP) feature extraction techniques in most cases. Statistical significance tests reveal that the improvement is most significant in high noise conditions. The technique thus provides improved robustness to noise without sacrificing performance in clean conditions     

Enhanced grasshopper optimization algorithm with extreme learning machines for motor-imagery classification

In Brain Computer Interface (BCI), achieving a reliable motor-imagery classification is a challenging task. The set of discriminative and relevant feature vectors plays a crucial role in classification. In this article, an enhanced optimization technique is implemented for selecting active feature vectors to enhance motor-imagery classification using Electroencephalography (EEG) signals. After collecting the input EEG signals from BCI competition III-4a and IV-2a databases, the 6th-order butter-worth filter is employed for eliminating base-line wander noise from the raw EEG signals. Further, the Variational Mode Decomposition technique is applied for separating the important signal components from the composite EEG signals, and then, the Higher Order Statistic, kurtosis, skewness, standard deviation, and entropy are utilized for feature extraction. The high-dimensional feature values are given to the Enhanced Grasshopper Optimization Algorithm for optimum feature selection, which are given to the Extreme Learning Machines (ELM) classifier for motor-imagery classification. Finally, in the resulting section, the optimized ELM model achieved 99.48% and 99.12% of accuracy on the BCI competition III-4a and IV-2a databases, where the achieved results are maximum compared to the traditional deep learning models.     

基于EMD和PCA的P300分类算法

提出一种基于经验模态分解(EMD)及主分量分析(PCA)的分类算法,采用支持向量机(SVM)对P300脑电信号字符拼写实验进行分类,通过EMD变换对P300脑电信号分解,从而达到去噪增强特征的效果,使用PCA方法对原始P300信号进行特征提取和集中,并送入SVM中实现分类。实验结果表明,该算法能获得高达96%的分类正确率。     

基于EMD和PCA的P300分类算法

提出一种基于经验模态分解(EMD)及主分量分析(PCA)的分类算法,采用支持向量机(SVM)对P300脑电信号字符拼写实验进行分类,通过EMD变换对P300脑电信号分解,从而达到去噪增强特征的效果,使用PCA方法对原始P300信号进行特征提取和集中,并送入SVM中实现分类。实验结果表明,该算法能获得高达96%的分类正确率。     

Analysis of EEG signals by implementing eigenvector methods/recurrent neural networks

The implementation of recurrent neural network (RNN) employing eigenvector methods is presented for classification of electroencephalogram (EEG) signals. In practical applications of pattern recognition, there are often diverse features extracted from raw data which needs recognizing. Because of the importance of making the right decision, the present work is carried out for searching better classification procedures for the EEG signals. Decision making was performed in two stages: feature extraction by eigenvector methods and classification using the classifiers trained on the extracted features. The aim of the study is classification of the EEG signals by the combination of eigenvector methods and the RNN. The present research demonstrated that the power levels of the power spectral density (PSD) estimates obtained by the eigenvector methods are the features which well represent the EEG signals and the RNN trained on these features achieved high classification accuracies.     

An efficient classifier to diagnose of schizophrenia based on the EEG signals

In this paper, electroencephalogram (EEG) signals of 13 schizophrenic patients and 18 age-matched control participants are analyzed with the objective of classifying the two groups. For each case, multi-channels (22 electrodes) scalp EEG is recorded. Several features including autoregressive (AR) model parameters, band power and fractal dimension are extracted from the recorded signals. Leave-one (participant)-out cross validation is used to have an accurate estimation for the separability of the two groups. Boosted version of Direct Linear Discriminant Analysis (BDLDA) is selected as an efficient classifier which applied on the extracted features. To have comparison, classifiers such as standard LDA, Adaboost, support vector machine (SVM), and fuzzy SVM (FSVM) are applied on the features. Results show that the BDLDA is more discriminative than others such that their classification rates are reported 87.51%, 85.36% and 85.41% for the BDLDA, LDA, Adaboost, respectively. Results of SVM and FSVM classifiers were lower than 50% accuracy because they are more sensitive to outlier instances. In order to determine robustness of the suggested classifier, noises with different amplitudes are added to the test feature vectors and robustness of the BDLDA was higher than the other compared classifiers.