Segmentation and tracking of object when grasped and moved within living spaces.
Takuya OmiKoh KakushoMasaaki IiyamaSatoshi NishiguchiPublished in: SMC (2017)
Keyphrases
- multiple objects
- object segmentation
- object tracking
- cluttered background
- deformable objects
- bounding box
- particle filter
- occlusion reasoning
- object matching
- target object
- motion segmentation
- motion cues
- d objects
- segmentation algorithm
- object region
- object position
- tracking objects
- textured objects
- pixel level
- rigid objects
- signed distance
- image segmentation
- real time
- pose recovery
- visual tracking
- tracking framework
- automatic initialization
- image segments
- global shape
- accurate localization
- segmentation method
- video objects
- articulated objects
- image regions
- multiscale
- partial occlusion
- kalman filter
- level set
- medical images
- object model
- moving objects
- complex background
- image sequences
- video sequences
- detecting objects
- region growing
- mean shift
- data association
- detection responses
- partially occluded
- energy function
- object motion
- shape prior
- video surveillance