Graph-CSPNet: A Graph-Based Common Spatial Pattern Network for Motor Imagery Classification (2023)
Open in web
Code: https://github.com/GeometricBCI/Tensor-CSPNet-and-Graph-CSPNet
1 Abstract
The graph-based common spatial pattern network (Graph-CSPNet) extends the geometric deep learning framework Tensor-CSPNet by structuring the temporally and spectrally segmented EEG covariance features as graph nodes and learning on the SPD manifold with manifold-valued graph convolutions for motor imagery classification.
2 NOTES
Not in Zotero (arXiv only); local text extraction stored at /tmp/opencode/drn-pdfs2/GraphCSPNet_Ju2023.txt.
Relevant Topics:
Quotes:
Quote
Quote here.