Spectral Shrinkage of Tyler's $M$-Estimator of Covariance Matrix.
Arnaud BreloyEsa OllilaFrédéric PascalPublished in: CAMSAP (2019)
Keyphrases
- covariance matrix
- covariance matrices
- laplacian matrix
- principal component analysis
- sample size
- objective function
- multivariate gaussian
- mahalanobis distance
- denoising
- eigenvalues and eigenvectors
- gaussian mixture
- positive definite
- pseudo inverse
- image denoising
- estimation error
- eigendecomposition
- image processing
- principal components
- computer vision
- symmetric matrix
- spectral analysis
- spectral methods
- correlation matrix
- maximum likelihood
- search algorithm
- geometrical interpretation