Paper Abstract and Keywords |
Presentation |
2019-10-27 09:20
Word Recognition using word likelihood vector from speech-imagery EEG Satoka Hirata, Yurie Iribe (Aichi Prefectual Univ.), Kentaro Fukai, Kouichi Katsurada (Tokyo Univ. of Science), Tsuneo Nitta (Waseda Univ./Toyohashi Univ. of Tech.) SP2019-29 WIT2019-28 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Previous research suggests that humans manipulate the machine using their electroencephalogram called BCI (Brain Computer Interface). However, there are not existed effective methods for speech imagery recognition. In this report, we performed word recognition of 10 numbers using speech-imagery EEG signals that could be measured through a non-invasive technique to reduce user’s burden. After removing the noise contained in EEG signals, line spectra feature was extracted from the amplitude spectra obtained by LPA. Next, word likelihood vector was calculated based on line spectra feature. We carried out the word recognition experiment of 10 numbers using one or the other or both of line spectra and word likelihood vector to clear the effective features. As a result, it was clarified that the likelihood vector based method got the highest accuracy in all features. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Speech-imagery EEG / Word recognition / Likelihood vector / / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 119, no. 250, SP2019-29, pp. 69-73, Oct. 2019. |
Paper # |
SP2019-29 |
Date of Issue |
2019-10-19 (SP, WIT) |
ISSN |
Print edition: ISSN 0913-5685 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 |
SP2019-29 WIT2019-28 |
|