Approximation-based Common Principal Component for feature extraction in multi-class Brain-Computer Interfaces

Dat Tran, Xu Huang

Research output: A Conference proceeding or a Chapter in BookConference contribution

Abstract

Common Spatial Pattern (CSP) is a state-of-the-art method for feature extraction in Brain-Computer Interface (BCI) systems. However it is designed for 2-class BCI classification problems. Current extensions of this method to multiple classes based on subspace union and covariance matrix similarity do not provide a high performance. This paper presents a new approach to solving multi-class BCI classification problems by forming a subspace resembled from original subspaces and the proposed method for this approach is called Approximation-based Common Principal Component (ACPC). We perform experiments on Dataset 2a used in BCI Competition IV to evaluate the proposed method. This dataset was designed for motor imagery classification with 4 classes. Preliminary experiments show that the proposed ACPC feature extraction method when combining with Support Vector Machines outperforms CSP-based feature extraction methods on the experimental dataset
Original languageEnglish
Title of host publicationProceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS
EditorsJessica M. Lotito
Place of PublicationLondon
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages5061-5064
Number of pages4
Volume40
ISBN (Electronic)9781457702167
DOIs
Publication statusPublished - 2013
Event2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) - Osaka, Osaka, Japan
Duration: 3 Jul 20137 Jul 2013

Conference

Conference2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC)
Abbreviated titleEMBC 13
CountryJapan
CityOsaka
Period3/07/137/07/13

Fingerprint

Brain computer interface
Feature extraction
Covariance matrix
Support vector machines
Experiments

Cite this

Tran, D., & Huang, X. (2013). Approximation-based Common Principal Component for feature extraction in multi-class Brain-Computer Interfaces. In J. M. Lotito (Ed.), Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS (Vol. 40, pp. 5061-5064). London: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/EMBC.2013.6610686
Tran, Dat ; Huang, Xu. / Approximation-based Common Principal Component for feature extraction in multi-class Brain-Computer Interfaces. Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS. editor / Jessica M. Lotito. Vol. 40 London : IEEE, Institute of Electrical and Electronics Engineers, 2013. pp. 5061-5064
@inproceedings{d716ef84a2de48de8ab15729c10688a9,
title = "Approximation-based Common Principal Component for feature extraction in multi-class Brain-Computer Interfaces",
abstract = "Common Spatial Pattern (CSP) is a state-of-the-art method for feature extraction in Brain-Computer Interface (BCI) systems. However it is designed for 2-class BCI classification problems. Current extensions of this method to multiple classes based on subspace union and covariance matrix similarity do not provide a high performance. This paper presents a new approach to solving multi-class BCI classification problems by forming a subspace resembled from original subspaces and the proposed method for this approach is called Approximation-based Common Principal Component (ACPC). We perform experiments on Dataset 2a used in BCI Competition IV to evaluate the proposed method. This dataset was designed for motor imagery classification with 4 classes. Preliminary experiments show that the proposed ACPC feature extraction method when combining with Support Vector Machines outperforms CSP-based feature extraction methods on the experimental dataset",
keywords = "Brain-computer interface, Common Principal Component, Feature Extraction",
author = "Dat Tran and Xu Huang",
year = "2013",
doi = "10.1109/EMBC.2013.6610686",
language = "English",
volume = "40",
pages = "5061--5064",
editor = "Lotito, {Jessica M.}",
booktitle = "Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States",

}

Tran, D & Huang, X 2013, Approximation-based Common Principal Component for feature extraction in multi-class Brain-Computer Interfaces. in JM Lotito (ed.), Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS. vol. 40, IEEE, Institute of Electrical and Electronics Engineers, London, pp. 5061-5064, 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Osaka, Japan, 3/07/13. https://doi.org/10.1109/EMBC.2013.6610686

Approximation-based Common Principal Component for feature extraction in multi-class Brain-Computer Interfaces. / Tran, Dat; Huang, Xu.

Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS. ed. / Jessica M. Lotito. Vol. 40 London : IEEE, Institute of Electrical and Electronics Engineers, 2013. p. 5061-5064.

Research output: A Conference proceeding or a Chapter in BookConference contribution

TY - GEN

T1 - Approximation-based Common Principal Component for feature extraction in multi-class Brain-Computer Interfaces

AU - Tran, Dat

AU - Huang, Xu

PY - 2013

Y1 - 2013

N2 - Common Spatial Pattern (CSP) is a state-of-the-art method for feature extraction in Brain-Computer Interface (BCI) systems. However it is designed for 2-class BCI classification problems. Current extensions of this method to multiple classes based on subspace union and covariance matrix similarity do not provide a high performance. This paper presents a new approach to solving multi-class BCI classification problems by forming a subspace resembled from original subspaces and the proposed method for this approach is called Approximation-based Common Principal Component (ACPC). We perform experiments on Dataset 2a used in BCI Competition IV to evaluate the proposed method. This dataset was designed for motor imagery classification with 4 classes. Preliminary experiments show that the proposed ACPC feature extraction method when combining with Support Vector Machines outperforms CSP-based feature extraction methods on the experimental dataset

AB - Common Spatial Pattern (CSP) is a state-of-the-art method for feature extraction in Brain-Computer Interface (BCI) systems. However it is designed for 2-class BCI classification problems. Current extensions of this method to multiple classes based on subspace union and covariance matrix similarity do not provide a high performance. This paper presents a new approach to solving multi-class BCI classification problems by forming a subspace resembled from original subspaces and the proposed method for this approach is called Approximation-based Common Principal Component (ACPC). We perform experiments on Dataset 2a used in BCI Competition IV to evaluate the proposed method. This dataset was designed for motor imagery classification with 4 classes. Preliminary experiments show that the proposed ACPC feature extraction method when combining with Support Vector Machines outperforms CSP-based feature extraction methods on the experimental dataset

KW - Brain-computer interface

KW - Common Principal Component

KW - Feature Extraction

U2 - 10.1109/EMBC.2013.6610686

DO - 10.1109/EMBC.2013.6610686

M3 - Conference contribution

VL - 40

SP - 5061

EP - 5064

BT - Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS

A2 - Lotito, Jessica M.

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - London

ER -

Tran D, Huang X. Approximation-based Common Principal Component for feature extraction in multi-class Brain-Computer Interfaces. In Lotito JM, editor, Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS. Vol. 40. London: IEEE, Institute of Electrical and Electronics Engineers. 2013. p. 5061-5064 https://doi.org/10.1109/EMBC.2013.6610686