Escaping local minima through hierarchical model selection: Automatic object discovery, segmentation, and tracking in video.
Nebojsa JojicJohn M. WinnC. Lawrence ZitnickPublished in: CVPR (1) (2006)
Keyphrases
- model selection
- motion segmentation
- object segmentation and tracking
- cross validation
- objects in video sequences
- hyperparameters
- bayesian learning
- bayesian information criterion
- sample size
- parameter estimation
- video sequences
- regression model
- error estimation
- statistical inference
- mixture model
- generalization error
- machine learning
- feature selection
- gaussian process
- image segmentation
- model selection criteria
- level set
- selection criterion
- active contours
- multiple objects
- variable selection
- information criterion
- video frames
- motion model
- video data
- segmentation algorithm
- sobolev active contours
- parameter determination
- kalman filter
- visual tracking
- segmentation method
- particle filter
- feature vectors
- decision trees
- learning algorithm