Title:
RIEMANNIAN-SPACE-BASED METHOD FOR MIGRATING ELECTROENCEPHALOGRAM DATA OF NATURAL HAND MOVEMENT
Document Type and Number:
WIPO Patent Application WO/2024/055385
Kind Code:
A1
Abstract:
A Riemannian-space-based method for migrating electroencephalogram data of a natural hand movement. The method comprises: respectively extracting spatial covariance matrices of electroencephalogram signals of natural hand movements of a new user and historical users as a target data set and historical data sets; selecting a source data set from among the historical data sets on the basis of the sum of distances between Riemannian centers of different categories, and aligning Riemannian centers of the source data set and the target data set; retrieving source data set samples by using a distance relationship between the source data set samples and target data set samples, and adding same to a training set; performing geometric transformation on the aligned target data set in a Riemannian space; and finally, merging transformed target data set samples, which have labels, with the retrieved source data set samples to form a training set for the establishment of a model of the new user. The method can assist in the establishment of a model of a new user with the help of electroencephalogram data of natural hand movements of historical users, thereby shortening the time for model calibration.
Inventors:
XU BAOGUO (CN)
XUE MUHUI (CN)
PING JINGYU (CN)
DENG LEYING (CN)
SONG AIGUO (CN)
XUE MUHUI (CN)
PING JINGYU (CN)
DENG LEYING (CN)
SONG AIGUO (CN)
Application Number:
PCT/CN2022/126706
Publication Date:
March 21, 2024
Filing Date:
October 21, 2022
Export Citation:
Assignee:
UNIV SOUTHEAST (CN)
International Classes:
G06F18/00; G06N20/00
Foreign References:
| CN114548166A | 2022-05-27 | |||
| CN113705645A | 2021-11-26 | |||
| JP2020038557A | 2020-03-12 |
Other References:
WANG FEI; PING JINGYU; XU ZONGFENG; BI JINYING: "Classification of motor imagery using multisource joint transfer learning", REVIEW OF SCIENTIFIC INSTRUMENTS, AMERICAN INSTITUTE OF PHYSICS, 2 HUNTINGTON QUADRANGLE, MELVILLE, NY 11747, vol. 92, no. 9, 27 September 2021 (2021-09-27), 2 Huntington Quadrangle, Melville, NY 11747, XP012259867, ISSN: 0034-6748, DOI: 10.1063/5.0054912
LI, YAN: "Study on Transfer Learning-based Motor Imagery Brain-computer Interfaces", MASTER'S THESIS, 27 May 2021 (2021-05-27), CN, pages 1 - 57, XP009553066, DOI: 10.27232/d.cnki.gnchu.2021.001222
YANG, FEIYU: "Research on Motor Imagery EEG Signal Classification Algorithm Based on Transfer Learning", MASTER'S THESIS, 19 June 2020 (2020-06-19), CN, pages 1 - 75, XP009553067, DOI: 10.27151/d.cnki.ghnlu.2020.001316
LI, YAN: "Study on Transfer Learning-based Motor Imagery Brain-computer Interfaces", MASTER'S THESIS, 27 May 2021 (2021-05-27), CN, pages 1 - 57, XP009553066, DOI: 10.27232/d.cnki.gnchu.2021.001222
YANG, FEIYU: "Research on Motor Imagery EEG Signal Classification Algorithm Based on Transfer Learning", MASTER'S THESIS, 19 June 2020 (2020-06-19), CN, pages 1 - 75, XP009553067, DOI: 10.27151/d.cnki.ghnlu.2020.001316
Attorney, Agent or Firm:
NANJING ZHONGLIAN PATENT AGENCY CO., LTD. (CN)
Download PDF:
Previous Patent: MULTIFUNCTIONAL SOUND PRODUCTION DEVICE
Next Patent: RESPIRATORY FEATURE EXTRACTION METHOD BASED ON BODY SURFACE SALIENCY ANALYSIS
Next Patent: RESPIRATORY FEATURE EXTRACTION METHOD BASED ON BODY SURFACE SALIENCY ANALYSIS