Recognition and tracking of object positions in image sequences using multiple-viewpoint television broadcast images.
Kazunori SakakiKenichiro YokotaKeiji OhnoKohei SuzukiHiroshi ArisawaPublished in: Systems and Computers in Japan (2007)
Keyphrases
- partial occlusion
- multiple objects
- objects in cluttered scenes
- viewpoint
- image sequences
- object tracking
- lighting conditions
- deformable objects
- target object
- object recognition
- illumination conditions
- three dimensional
- d objects
- object matching
- active object recognition
- rigid objects
- polyhedral objects
- object motion
- image frames
- three dimensional objects
- rigid body
- video images
- real world scenes
- particle filter
- robust tracking
- motion analysis
- object models
- image matching
- cluttered background
- input image
- textured objects
- calibrated cameras
- bounding box
- image features
- appearance model
- image retrieval
- image regions
- visual tracking
- object appearance
- traffic signs
- moving objects
- camera views
- object shape
- articulated objects
- object class
- motion segmentation
- mean shift
- viewing direction
- moving camera
- keypoints
- human gait
- camera movement
- video sequences
- multiple views
- human motion
- digital television
- background subtraction