Bayesian Method for Motion Segmentation and Tracking in Compressed Videos.
Siripong TreetasanatavornUwe RauschenbachJörg HeuerAndré KaupPublished in: DAGM-Symposium (2005)
Keyphrases
- state space
- motion segmentation
- bayesian methods
- model selection
- dynamic scenes
- motion model
- particle filter
- hyperparameters
- feature trajectories
- image sequences
- moving objects
- motion estimation
- missing data
- multibody
- feature tracking
- motion field
- gaussian processes
- cross validation
- multiple moving objects
- moving camera
- video sequences
- video surveillance
- optical flow
- bayesian framework
- machine learning
- feature selection
- video frames
- gaussian process
- spectral clustering
- active learning
- image processing
- image segmentation
- learning algorithm
- machine learning methods
- training data