A Novel Trajectory Clustering Approach for Motion Segmentation.
Matthias ZeppelzauerMaia ZaharievaDalibor MitrovicChristian BreitenederPublished in: MMM (2010)
Keyphrases
- motion segmentation
- spectral clustering
- feature trajectories
- missing data
- motion estimation
- image sequences
- motion model
- multibody
- moving objects
- model selection
- optical flow
- clustering method
- clustering algorithm
- point trajectories
- data clustering
- k means
- flow field
- feature tracking
- motion field
- dynamic scenes
- epipolar geometry estimation
- motion sequences
- unsupervised learning
- machine learning
- multiple moving objects
- depth ordering
- data points
- data mining