A new EM algorithm using Tikhonov regularization.
Takamasa KoshizenYves RosseelYoshihiro TonegawaPublished in: IJCNN (1999)
Keyphrases
- em algorithm
- tikhonov regularization
- regularization parameter
- hyperparameters
- image restoration
- expectation maximization
- reproducing kernel hilbert space
- mixture model
- maximum likelihood
- generative model
- parameter estimation
- gaussian mixture model
- density estimation
- incomplete data
- probability density function
- least squares
- bayesian framework
- gaussian process
- machine learning
- maximum a posteriori
- kernel methods
- unsupervised learning
- kernel matrix
- probabilistic model
- image segmentation