Circularly establishing feature correspondences and motion estimation for object tracking via multivisit of Kalman filtering.
Jean GaoAkio KosakaPublished in: Image Processing: Algorithms and Systems (2002)
Keyphrases
- object tracking
- kalman filtering
- kalman filter
- motion parameters
- motion estimation
- image sequences
- video sequences
- particle filtering
- multi frame
- structure from motion
- particle filter
- moving objects
- computer vision
- mean shift
- appearance model
- visual tracking
- camera motion
- motion model
- bundle adjustment
- optical flow
- multiple cameras
- multiple views
- multi camera
- motion vectors
- motion compensation
- sparse representation
- motion field
- motion segmentation
- motion analysis
- training set
- flow field
- feature points
- super resolution
- multiple images
- epipolar geometry
- optic flow
- input image
- computational complexity