An Infinitesimal Probabilistic Model for Principal Component Analysis of Manifold Valued Data.
Stefan SommerPublished in: CoRR (2018)
Keyphrases
- valued data
- probabilistic model
- principal component analysis
- low dimensional
- mixture model
- covariance matrix
- covariance matrices
- riemannian metric
- feature space
- tensor field
- dimensionality reduction
- principal components
- order statistics
- generative model
- matrix valued
- bayesian networks
- face recognition
- language model
- riemannian manifolds
- gaussian mixture model
- expectation maximization
- face images
- high dimensional data
- feature extraction
- high dimensional
- data points
- image processing