Paper Abstract and Keywords |
Presentation |
2019-03-05 16:15
Performance improvement of fNIRS-BCI using generative adversarial networks Tomoyuki Nagasawa (Nagaoka Univ. of Tech.), Takanori Sato (National Inst. of Tech., Akita Col.), Isao Nambu, Yasuhiro Wada (Nagaoka Univ. of Tech.) NC2018-72 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Since a lengthy functional near-infrared spectroscopy (fNIRS) measurement is uncomfortable for the participant, the number of data that can be acquired is limited. As a result, the training data of the classification model is insufficient; hence, the brain-computer interface (BCI) performance decreases. In this study, to improve the fNIRS-BCI accuracy, we examined an fNIRS data-augmentation method using generative adversarial networks (GANs). When applying the proposed method to the 4-class classification of a left-, right-, or both-handed ball grasping or resting, the accuracy improved after the data augmentation. This result suggests that data augmentation using GANs is useful for improving the fNIRS-BCI performance. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
generative adversarial networks / data augmentation / functional near-infrared spectroscopy / brain-computer interface / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 118, no. 470, NC2018-72, pp. 151-156, March 2019. |
Paper # |
NC2018-72 |
Date of Issue |
2019-02-25 (NC) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
NC2018-72 |
|