Maximum Likelihood Estimates for Gaussian Mixtures Are Transcendental.
Carlos AméndolaMathias DrtonBernd SturmfelsPublished in: MACIS (2015)
Keyphrases
- gaussian mixture
- maximum likelihood estimates
- em algorithm
- closed form
- expectation maximization
- maximum likelihood
- parameter estimation
- likelihood function
- mixture model
- gaussian mixture model
- hyperparameters
- point correspondences
- probability density function
- generative model
- probabilistic model
- motion parameters
- linear models
- image reconstruction
- bayesian framework
- maximum a posteriori
- machine learning
- unsupervised learning
- k means
- object recognition
- image sequences
- image segmentation