Use of the mean-field approximation in an EM-based approach to unsupervised stochastic model-based image segmentation.
David A. LanganKarl J. MolnarJames W. ModestinoJun ZhangPublished in: ICASSP (1992)
Keyphrases
- unsupervised learning
- expectation maximization
- monte carlo
- em algorithm
- completely unsupervised
- maximum likelihood
- probabilistic model
- finite mixture model
- generative model
- case based reasoning
- stochastic model
- mixture model
- image segmentation
- graphical models
- semi supervised
- object recognition
- data driven
- image processing
- real time
- gaussian mixture model
- density estimation
- supervised classification
- natural language processing
- stochastic process
- stochastic optimization
- stochastic nature
- database