Title: Eigenface analysis for brain signal classification: a novel algorithm

Authors: Yeon-Mo Yang; Wansu Lim; Byeong Man Kim

Addresses: School of Electronic Engineering, Kumoh National Institute of Technology, Gumi, 39177, South Korea ' Department of IT Convergence, Kumoh National Institute of Technology, Gumi, 39177, South Korea ' Department of Computer Software Engineering, Kumoh National Institute of Technology, Gumi, 39177, South Korea

Abstract: This paper proposes a novel feature extraction scheme utilising an Eigenface analysis (EFA) algorithm for a brain computer interface (BCI). In EFA, the obtained BCI data is systematically rearranged into time, channels, and trials to develop neuro-images. Based on these images, the scheme extracts Eigenfaces with a training dataset and utilises the cross-correlation to find the coefficients of projection. Compared to the existing scheme, EFA outperforms in accuracy with BCI competition III, dataset IIIa. Specifically, the accuracy improves by 27.21% for the second subject.

Keywords: bio medical signal processing; neuro data classification; hypothesis test; statistical signals; brain computer interface; BCI; Eigen vector decomposition.

DOI: 10.1504/IJTMCP.2017.083887

International Journal of Telemedicine and Clinical Practices, 2017 Vol.2 No.2, pp.148 - 153

Received: 18 Oct 2016
Accepted: 11 Nov 2016

Published online: 25 Apr 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article