Transfer Learning: A Riemannian Geometry Framework With Applications to Brain–Computer Interfaces

Hence, we make, through the affine transformation proposed, data from different sessions and subject comparable, providing a significant improvement in the BCI transfer learning problem.