Sparse Coding on Symmetric Positive Definite Manifolds Using Bregman Divergences.
Mehrtash Tafazzoli HarandiRichard I. HartleyBrian C. LovellConrad SandersonPublished in: IEEE Trans. Neural Networks Learn. Syst. (2016)
Keyphrases
- sparse representation
- riemannian manifolds
- manifold learning
- euclidean space
- natural images
- image classification
- feature space
- positive definite
- image patches
- image representation
- kl divergence
- high dimensional data
- face recognition
- geometric structure
- vector space
- dimensionality reduction
- loss function
- low dimensional
- geodesic distance
- high dimensional
- multiscale
- diffusion tensor
- unsupervised learning
- information theoretic
- mean shift
- covariance matrix
- metric space
- feature selection
- object recognition
- generative model
- similarity measure
- higher order
- probabilistic model
- learning theory
- multi class