| WANG Xiaoling (王晓玲),PANG Yu,HAN Changqing,ZHAO Ze,GAO Nuo.[J].高技术通讯(英文),2025,31(4):397~406 |
|
| Motor imagery EEG signal classification based on multi-riemannian kernel fusion features |
| |
| DOI：10. 3772 / j. issn. 1006-6748. 2025. 04. 009 |
| 中文关键词: |
| 英文关键词: motor imagery electroencephalogram signal, brain-computer interface, symmetric positive definite manifold, Gaussian symmetric positive definite manifold, Grassmann manifold |
| 基金项目: |
| Author Name | Affiliation | | WANG Xiaoling (王晓玲) | (College of Information and Electrical Engineering, Shandong Jianzhu University, Jinan 250101, P. R. China) | | PANG Yu | | | HAN Changqing | | | ZHAO Ze | | | GAO Nuo | |
|
| Hits: 9 |
| Download times: 8 |
| 中文摘要: |
| |
| 英文摘要: |
| The classification of motor imagery electroencephalogram (MI-EEG) signals is one of the key challenges in brain-computer interface ( BCI ) technology. Existing Riemannian geometry-based methods for MI-EEG signal analysis, which rely on a single symmetric positive definite (SPD) manifold, often provide a limited geometric structure, making it difficult to fully capture the complex geometric characteristics of the signals. To address this issue, this paper proposes an innovative classification method for MI-EEG signals based on multi-Riemannian kernel fusion features (MRKFF).This method extends the classical SPD manifold by incorporating the Gaussian SPD manifold and the Grassmann manifold, extracting more discriminative kernel features from these heterogeneous manifolds for fusion-based classification. The proposed method is validated on the OpenBMI binary classification dataset and the BCI Competition IV-2a four-class dataset, achieving average classification accuracies of 75. 6% and 71. 0% , with Kappa values of 0. 50 and 0. 61, respectively. The proposed MRKFF method provides a new perspective for the geometric analysis of MI-EEG signals, enabling a deeper understanding and analysis of the complex geometric structure of these signals, thereby achieving more accurate signal classification in BCI applications. |
|
View Full Text
View/Add Comment Download reader |
| Close |
|
|
|
