Research on Feature Recognition on mVEP BCI

DOI

10.2991/icmeit-17.2017.29How to use a DOI?

Keywords

locally liner embedding; kernel entropy component analysis; kernel entropy liner embedding.

Abstract

The recognition to element N2 is the fundamental basis determining the stability and the practicability of the mVEP-BCI systems. The feature extraction to EEG signals is the key for the accuracy on recognition of element N2. Different from the traditional down sampling feature extraction method, the method in this article is designed by utilizing the over-complete dictionary based compressed sensing method, to conduct dimension reduction processing to EEG signals in the time window by utilizing the row echelon observation matrix of different compression ratios for multiple times, to acquire the ordinary features of mVEP signals. It further conducts sparse noise reduction to ordinary features. It also conducts LDA classifications on down sampling features and the features extracted in this article. According to verification, the classification accuracy rate of the features extracted in this article has significant improvement than that of the traditional method. The feature extraction method in this article improves classification effect by taking the avoidance of overfitting and the retaining of useful information of original signal to the maximum extent into consideration, which improves the classification effect and enhances the stability and practicability of mVEP-BCI system in a more efficient way.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Teng Ma
AU  - Hui Li
AU  - Dezhong Yao
AU  - Peng Xua
PY  - 2017/05
DA  - 2017/05
TI  - Research on Feature Recognition on mVEP BCI
BT  - Proceedings of the 2nd International Conference on Mechatronics Engineering and Information Technology (ICMEIT 2017)
PB  - Atlantis Press
SP  - 160
EP  - 165
SN  - 2352-538X
UR  - https://doi.org/10.2991/icmeit-17.2017.29
DO  - 10.2991/icmeit-17.2017.29
ID  - Ma2017/05
ER  -