Title :
Real-time patternbases for multiple targets in gaze control for target tracking systems
Author :
Ramanna, Sheela ; Baumela, Luis
Author_Institution :
Dept. of Comput. Sci., Hartford Univ., West Hartford, CT, USA
Abstract :
A method for selecting a target among multiple targets in terms of the use of a patternbase is presented. A many-to-one target methodology has been chosen, since we use a patternbase to construct internal models for possible objects in the field of vision of a camera based target tracking system. The patternbase approach to gaze control presented introduces collections of (attribute, value) pairs which characterize a possible target. Each collection of (attribute, value) pairs forms a patternbase which represents an internal model useful in the identification tasking for a target tracking system. As a result, the corresponding patternbase would be expanded dynamically to accommodate the changing characteristics of an object. The contribution of the paper is the design of real time patternbase management system useful in gaze control for target tracking in the presence of multiple targets
Keywords :
active vision; pattern recognition; real-time systems; robot vision; target tracking; tracking; active vision; camera based target tracking system; gaze control; internal model; internal models; many-to-one target methodology; multiple targets; real time patternbase management system; target tracking systems; Cameras; Computed tomography; Computer science; Control systems; Lenses; Mobile robots; Real time systems; Robot vision systems; Target tracking; Visual system;
Conference_Titel :
WESCANEX 95. Communications, Power, and Computing. Conference Proceedings., IEEE
Conference_Location :
Winnipeg, Man.
Print_ISBN :
0-7803-2725-X
DOI :
10.1109/WESCAN.1995.493950