Distance based kernels for video tensors on product of Riemannian matrix manifolds.
Krishan SharmaRenu RameshanPublished in: J. Vis. Commun. Image Represent. (2021)
Keyphrases
- symmetric positive definite matrices
- matrix valued
- riemannian manifolds
- lie group
- symmetric positive definite
- positive definite
- log euclidean
- dt mri
- euclidean space
- tensor field
- vector valued
- valued data
- fisher information
- manifold learning
- vector space
- differential geometry
- diffusion tensor
- information geometry
- video data
- riemannian metric
- geometric structure
- video sequences
- parameter space
- euclidean distance
- distance measure
- covariance matrix
- video frames
- multimedia
- feature space
- kernel function
- square root
- vector field
- geodesic paths
- riemannian framework
- multiple kernel learning
- low dimensional
- covariance matrices
- high order
- geodesic distance
- high dimensional
- positive semidefinite
- linear combination
- mean shift
- structure tensor
- dimensionality reduction
- data points
- reproducing kernel hilbert space
- visual data