MulticlassCommonSpatialPatternfor_共空间模式资源-CSDN文库

需积分: 9 80 浏览量 2020-03-19 16:17:32 上传评论收藏 497KB PDF 举报

标题中的“Multiclass Common Spatial Pattern”表明文章内容主要围绕多类公共空间模式（CSP）算法在基于脑电图（EEG）的脑机接口（BCI）中的应用。CSP算法是一种用于分析多通道信号数据，特别是脑电波数据的技术，它通过空间滤波器提取相关信号特征，从而提高信号的分类准确率。该算法在运动想象任务中表现出有效性，能够提取出有助于分类的特征。描述中提到，CSP算法虽然在脑机接口中有效，但容易过拟合，尤其是当问题扩展到多类别时，现有算法就不再那么有效了。此外，数据中的异常值和非平稳性也是CSP特征提取的一个挑战。作者提出了一种在预处理阶段识别并移除数据中的伪迹（artifacts）的方法，这有助于计算出更好的特征向量。为了处理特征的非平稳性问题，文章扩展了文献中提出的自调节区间型-2神经模糊推理系统（SRIT2NFIS）到多类别问题，并采用联合近似对角化（JAD）方法。在BCI竞赛IV的标准数据集上进行的实验显示，与当前最先进的方法相比，准确率有了显著提高。标签“CSP算法”表明，这是研究CSP算法相关知识的文献，强调算法在特定领域的应用，即脑机接口，以及如何优化其性能。部分内容表明，BCI的目标是建立大脑和计算机/机器之间的非肌肉通信方式，它通过解释大脑中的思维活动，将其转化为机器的输出。BCI是一个系统，它采集大脑信号活动并将其翻译成输出，从而能够替代、恢复、增强、补充或改善现有系统的功能。在具体实现CSP算法时，需要处理信号的非平稳性，以及信噪比低和伪迹污染等问题。通过预处理阶段移除伪迹可以改善特征向量的计算，从而生成更优质的CSP特征。而处理特征非平稳性的挑战，则通过采用自调节区间型-2神经模糊推理系统（SRIT2NFIS）来扩展到多类别问题，并采用联合近似对角化（JAD）方法来实现。这样处理后，在BCI竞赛IV的标准数据集上的实验结果显示出相比于当前最佳技术显著提高了准确率。

资源推荐

资源详情

资源评论

Pattern Recognition Letters

journal homepage: www.elsevier.com

Multiclass Common Spatial Pattern for EEG based Brain Computer Interface with

Adaptive Learning Classiﬁer

Hardik Meisheri

a,∗∗

, Nagaraj Ramrao

, Suman K. Mitra

Dhirubhai Ambani Institute of Information and Communication Technology, Indroda Circle, Gandhinagar - 382007, India

ABSTRACT

In Brain Computer Interface (BCI), data generated from Electroencephalogram (EEG) is non-station-

ary with low signal to noise ratio and contaminated with artifacts. Common Spatial Pattern (CSP)

algorithm has been proved to be eﬀective in BCI for extracting features in motor imagery tasks, but it

is prone to overﬁtting. Many algorithms have been devised to regularize CSP for two class problem,

however they have not been eﬀective when applied to multiclass CSP. Outliers present in data aﬀect

extracted CSP features and reduces performance of the system. In addition to this non-stationarity

present in the features extracted from the CSP present a challenge in classiﬁcation. We propose a

method to identify and remove artifact present in the data during pre-processing stage, this helps in

calculating eigenvectors which in turn generates better CSP features. To handle the non-stationarity,

Self-Regulated Interval Type-2 Neuro-Fuzzy Inference System (SRIT2NFIS) was proposed in the lit-

erature for two class EEG classiﬁcation problem. This paper extends the SRIT2NFIS to multiclass

using Joint Approximate Diagonalization (JAD). The results on standard data set from BCI compe-

tition IV shows signiﬁcant increase in the accuracies from the current state of the art methods for

multiclass classiﬁcation.

1. Introduction

There is on-going research to connect brain signals with

computers/machines in order to create a comfortable communi-

cation pathway for disabled people with external environment.

Brain Computer Interface (BCI) aims to provide non-muscular

communication between brain and computer/machine by inter-

preting the thoughts in the brain. BCI can be deﬁned as: ”a

system that acquires brain signal activity and translates it into

an output that can replace, restore, enhance, supplement, or im-

prove the existing brain signal, which can, in turn, modify or

change ongoing interactions between the brain and its internal

or external environment”. Or in simple words, it is a system that

translates brain signals into new kinds of outputs as mentioned

in Daly and Wolpaw (2008). In intelligent devices, which rely

hugely upon the user experience, BCI is used for monitoring

purposes, for example monitoring attention level of driver on a

long drive. Apart from clinical application for disabled people

∗∗

e-mail: hardik_meisheri@daiict.ac.in (Hardik Meisheri)

for controlling devices, BCI is used in virtual reality, gaming

industry and entertainment industry. It is also used for medita-

tion and rehabilitation purposes. BCI is employed in boosting

learning abilities of children who have diﬃculty with this.

There are three broad categories of BCIs namely, im-

plantable/invasive, non-invasive and partial/semi-invasive, dis-

tinguished by invasively and noninvasively acquired brain sig-

nals, respectively. Invasive BCI have a requirement of placing

the sensors inside the brain, for which operation is necessary.

Although this type of BCI provide one of the best signal-to-

noise ratio (SNR), they are not preferred due to high economic

cost, short sensor duration and surgical risks. Non-invasive BCI

can be realized using brain’s electric or magnetic ﬁeld, out of

which electric ﬁeld is preferred choice as it is mobile and set of

apparatus needed for this is easily available and not bulky. Elec-

troencephalography (EEG) has been used to detect the electric

signals in a non-invasive manner. BCI using EEG can be fur-

ther catergorized into event related potential - P300 proposed

by Farwell and Donchin (1988), steady state visual evoked po-

tential (SSVEP) mentioned in Vidal (1973) and event related

desynchronization/synchronization (ERD/S). In this paper, we

arXiv:1802.09046v1 [cs.NE] 25 Feb 2018

concentrate on motor imagery system based BCI which is a type

of ERD/S.

Motor Imagery based BCI were invented to help people with

disabilities to communicate with the outer world. EEG has

proven to be eﬀective in motor imagery based BCI due to its

very light equipment, low cost and high temporal resolution

Curran and Stokes (2003). However, BCI using EEG suﬀers

from challenges such as extracting features which are useful

for speciﬁc task due to low speciﬁcity, vulnerable to volume

conduction eﬀects, non-stationarity, and prone to noise Wol-

paw et al. (2000). Another problem posed by EEG signals is

that they vary from person to person and session to session

Krauledat et al. (2007). Spatial ﬁltering has been introduced

to discriminate between the motor imagery signals using multi-

channel EEG. The objective behind this ﬁltering is to transform

the multichannel EEG signals into small set of channels which

are useful to distinguish between the diﬀerent brain activities

Vidaurre et al. (2011); Tangwiriyasakul et al. (2013).

2. Related Work

Common spatial Pattern (CSP) has proven to be very eﬃ-

cient in context of motor imagery based BCI Tangermann et al.

(2012); Ramoser et al. (2000); Blankertz et al. (2008, 2004,

2006). Ideally, CSP computes the ﬁlters which maximizes the

ratio of variance between the brain activities, details of which

are explained in the subsection 3.1. However, CSP is vulnerable

to noise and problem of overﬁtting Reuderink and Poel (2008);

Grosse-Wentrup et al. (2009). To overcome these shortcom-

ings, many methods have been proposed in literature Blankertz

et al. (2008); Lu et al. (2009); Kang et al. (2009); Lotte and

Guan (2010). Lotte and Guan (2011) presents detailed com-

parisons of these regularization techniques. Dornhege et al.

(2004a) in their pioneer work extended CSP to multiclass by

using methods such as one-versus-rest, pair-wise and simulta-

neous diagonalization. A major drawback in extending CSP to

multiclass is that regularization methods proposed for two class

are not eﬀective, therefore accuracies for multiclass are sub-

stantially low. For extending original CSP algorithm to more

than two class, Naeem et al. (2006), Brunner et al. (2007),

Grosse-Wentrup and Buss (2008), Gouy-Pailler et al. (2008),

Gouy-Pailler et al. (2010), Grosse-Wentrup and Buss (2008)

and Wei et al. proposed diﬀerent methods using Joint Ap-

proximate Diagonalization (JAD) . Among all these approaches

common part was that they assumed that CSP is equivalent to

ﬁnding independent common components for 2-class problem.

Naeem et al. (2006) and Brunner et al. (2007) researched on dif-

ferent Independent Component Analysis (ICA) algorithms for

ﬁnding features and components including Informax Bell and

Sejnowski (1995), FastICA Hyvarinen (1999) and SOBI Be-

louchrani et al. (1997). They presented key diﬀerences between

the performance of ICA-based methods with other CSP meth-

ods which are variants of One-versus-the- Rest and Pair-Wise

Dornhege et al. (2004b). Their ﬁndings concluded that overall,

CSP methods perform better than ICA-based methods. In addi-

tion to this they also stated that among the ICA-based methods,

Infomax achieved better performance than others. CSP is iden-

tical to ICA was proven mathematically by Grosse-Wentrup

and Buss (2008). The authors proposed that CSP by JAD is

identical to ICA and also proposed information theoretic ap-

proach for feature extraction. Gouy-Pailler in their recent work

Gouy-Pailler et al. (2008), Gouy-Pailler et al. (2010) proposed

maximum likelihood method for ﬁnding spatial ﬁlters which

is extension of JAD method. They claimed that their method

is a neurophysiologically adapted version of JAD. They vali-

dated their approach on Dataset 2a of the BCI Competition IV

which has nine subjects(they have used only 8 subjects out of

9) with four motor-imagery tasks, and reported that their newly

proposed JAD method achieved a better classiﬁcation accuracy

than the CSPs method. They also claimed that CSP with JAD is

not better signiﬁcantly than CSP methods which are variant of

of One-versus-the- Rest and Pair-Wise as reported in the ﬁnd-

ings of Grosse-Wentrup and Buss (2008). In a separate work,

Wei et al. used quadratic optimization to ﬁnd common spa-

tial patterns for multi-class BCI problems. However, in this

case they conducted experiments on their own dataset instead

on more widely used datasets so it is diﬃcult to compare their

results with previously noted research results. Apart from these

two methods there is some interesting work done using sub-

space method by Ramoser et al. (2000), in which they propose

Union-based common principal components (UCPC) to create

a subspace for class of data from covariance matrix, and ﬁnally

the union of all the all subspaces is used as common principal

component. However, drawback of this method is that chosen

principal components may not contribute to some data classes

at all.

In summary, multi-class BCI systems can be broken down

into two main groups. The ﬁrst is based on extension of two

class methods that original CSP was designed to work upon, on

the other hand other groups deals with JAD. Key point of two

class CSP- based methods is that it breaks down the multi-class

problems into various classiﬁcation two-class problems. The

two popular methods are One-versus-the-Rest, and a combina-

tion of pairs of two-class classiﬁcation problems (Pair-Wise).

Each of these two methods has its own weakness. In the ﬁrst

method there is an assumption that covariance matrix all the

other classes are almost identical. However, this assumption

hardly stands true in real world data. In latter, there is no surety

of getting the CSPs that are optimal of diﬀerent pairs of classes.

This is similar to grouping common principal components of

pairs of classes and then ﬁnding CSPs.

Non-stationarity nature of data in EEG presents major prob-

lem, which is due to the variation in intra- and inter- session in

the same task. This results in very low accuracies in the clas-

siﬁcation phase, due to poor modeling of data. Recently fuzzy

inference systems were used in EEG which has proved to be ef-

fective for classiﬁcation purpose. As shown in Guler and Ubeyli

(2005); Coyle et al. (2009); Fabien et al. (2007); Subasi (2007);

Yang et al. (2014) type-1 inference systems have been em-

ployed for classiﬁcation of EEG signals. However, type-1 fuzzy

sets are ineﬃcient in handling non-stationarity of EEG signals

as pointed out by Herman et al. (2006). Type-2 fuzzy sets which

was proposed by Zadeh (1975) along with uncertainty frame-

work can be used to handle non-stationarity as shown in Mendel

and John (2002). Type-2 sets are computationally very expen-

剩余7页未读，继续阅读

评论收藏

内容反馈

zhoudapeng01

粉丝: 278
资源: 33

Multiclass Common Spatial Pattern for

SVM_For_multiclass

MulticlassCSP.m

Multiclass_Classification_machineryh2g_matlab_multiclass_classif

Support-Vectror-Machine.rar_LibSVM_libsvm multiclass_multiclass_

ppyoloe模型部署KeyError: 'multiclass-nms3-0.tmp-2.lod'解决

MSVM.zip_msvm_multiclass svm_online svm_svm multiclass

U-net_for_Multiclass_Segmentation

Multiclass_Regularized_LDA-master_分类算法_

multiclass.rar_ multiclass_matlab多分类_multiclass matlab_分类算法_多分类算

NN_multiclass_classification_Pytorch_classification_

Weakly supervised multiclass video segmentation

Multiclass boosting SVM using different texture features in HEp-2 cell staining pattern classification

td_one_vs_rest_multiclass_classifier.rar_One Vs Rest

Mining Potential Information for Multiclass Microarray Data Using Centroid-Based Dimension Reduction

svmlight_multiclass.rar_svm multi_svm multiclass windo_svm 多类_支持

CNN---multiclass-classification

国家开放大学计算机应用基础终结性考试（大作业）

离散数学知识点整理（超级全面详细！）

《科研伦理与学术规范》期末考试文档2（40题）

Word2Recite 桌面单词

2021全国及分省市县行政区划矢量图层shp文件.rar

38000词汇思维导图（1-50词根）β版.rar

MCGS组态精品版图库.zip

Revit 各版本官方族库及项目样板下载和安装方法，2016-2021族库离线包下载.rar

博士“申请-考核制”面试——英文提问问题/答案模板

最新资源