Simultaneous motion parameter estimation and image segmentation using the EM algorithm.
Kristine E. MatthewsNader M. NamaziPublished in: ICIP (1995)
Keyphrases
- parameter estimation
- em algorithm
- expectation maximization
- image segmentation
- maximum likelihood
- mixture model
- markov random field
- maximum likelihood estimation
- gaussian mixture model
- motion analysis
- generative model
- random fields
- parameter estimation algorithm
- probabilistic model
- hyperparameters
- probability density function
- expectation maximisation
- camera motion
- motion estimation
- optical flow
- maximum a posteriori
- gaussian mixture
- likelihood function
- gibbs sampling
- image processing
- unsupervised learning
- motion model
- parameter learning
- level set
- approximate inference
- computer vision
- image sequences
- density estimation
- posterior distribution
- moving objects
- higher order
- bayesian model selection
- unsupervised segmentation
- incomplete data
- graph cuts
- segmentation algorithm
- structure learning
- graphical models
- bayesian networks
- maximum likelihood estimates
- active contours
- bayesian framework