Abstract | ||
---|---|---|
The non-stationarity inherent across sessions recorded on different days poses a major challenge for practical electroencephalography (EEG)-based Brain Computer Interface (BCI) systems. To address this issue, the computational model trained using the training data needs to adapt to the data from the test sessions. In this paper, we propose a novel approach to compute the variations between labelled training data and a batch of unlabelled test data based on the geodesic-distance of the discriminative subspaces of EEG data on the Grassmann manifold. Subsequently, spatial filters can be updated and features that are invariant against such variations can be obtained using a subset of training data that is closer to the test data. Experimental results show that the proposed adaptation method yielded improvements in classification performance. |
Year | DOI | Venue |
---|---|---|
2014 | 10.1109/IJCNN.2014.6889686 | IJCNN |
Keywords | Field | DocType |
spatial filters,geodesic distance,electroencephalography,brain-computer interfaces,practical electroencephalography-based brain computer interface systems,medical signal processing,motor eeg classification,spatial filter adaptation,session recording,signal classification,labelled training data,grassmann manifold,computational model,unlabelled test data,brain computer interfaces,manifolds,training data,computational modeling | Computer science,Brain–computer interface,Artificial intelligence,Discriminative model,Electroencephalography,Spatial filter,Pattern recognition,Speech recognition,Linear subspace,Invariant (mathematics),Test data,Geodesic,Machine learning | Conference |
ISSN | Citations | PageRank |
2161-4393 | 0 | 0.34 |
References | Authors | |
12 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Xinyang Li | 1 | 19 | 4.13 |
Cuntai Guan | 2 | 1298 | 124.69 |
Kai Keng Ang | 3 | 804 | 64.19 |
Haihong Zhang | 4 | 687 | 57.92 |
Sim Heng Ong | 5 | 426 | 44.63 |