A Max-Product EM Algorithm for Reconstructing Markov-Tree Sparse Signals From Compressive Samples.
Zhao SongAleksandar DogandzicPublished in: IEEE Trans. Signal Process. (2013)
Keyphrases
- em algorithm
- markov tree
- compressive sensing
- max product
- loopy belief propagation
- expectation maximization
- mixture model
- maximum likelihood
- belief propagation
- belief functions
- sparse representation
- parameter estimation
- tree structure
- generative model
- gaussian mixture model
- incomplete data
- message passing
- training samples
- hyperparameters
- probability density function
- training set
- hidden variables
- approximate inference
- graph cuts
- graphical models
- object recognition
- unsupervised learning
- upper bound
- probabilistic model