Comparison of Different Classifiers for Biometric System Based on EEG Signals

Author

Jian-feng, Hu

Author_Institution

Inst. of Inf. Technol., Jiangxi Bluesky Univ., Nanchang, China

fYear

2010

fDate

24-25 July 2010

Firstpage

288

Lastpage

291

Abstract

Characteristics in EEG signals related to the motor imagery can be used to build up a biometric system. However, for the practical implementation of a biometric system, the classifier plays a crucial role. In this paper, I compared the performance of three different classifiers for the detection of the imagined movements in a group of subjects on the basis of EEG signals. The classifiers compared here were those based on Linear Discrimination Analysis (LDA), Artificial Neural Network (ANN) and Support Virtual Machine (SVM). Results show a better performance of the LDA classifier with the respect to the other classifiers.

Keywords

biometrics (access control); electroencephalography; medical signal processing; neural nets; support vector machines; ANN; EEG signals; LDA; SVM; artificial neural network; biometric system; linear discrimination analysis; support virtual machine; Artificial neural networks; Biometrics; Brain modeling; Electroencephalography; Support vector machines; Testing; Training; Biometric; Classifier; Electroencephalogram (EEG); Linear Discrimination Analysis (LDA); Neural network; Support Virtual Machine (SVM);

fLanguage

English

Publisher

ieee

Conference_Titel

Information Technology and Computer Science (ITCS), 2010 Second International Conference on

Conference_Location

Kiev

Print_ISBN

978-1-4244-7293-2

Electronic_ISBN

978-1-4244-7294-9

Type

conf

DOI

10.1109/ITCS.2010.77

Filename

5557128