Title :
Least Squares Fitting Articulated Objects
Author :
C. Gronwall;P. Anderson;F. Gustafsson
Author_Institution :
Dept. Laser System, FOI Linkoping, Sweden
fDate :
6/27/1905 12:00:00 AM
Abstract :
In safety and security applications, one issue is target recognition. In some recognition processes, for vehicle recognition, we must take into account that the target maybe articulated. Vehicles can easily change shape by opening of a door, adding of load etc.This change of shape can be moddled as an articulation. if the articluation of model the target also can be moddled, the mathcing with library modles can be improved. In this paper we propose a method for modular least squares fitting of two 3D point scatters with points correspondence. A method for least squares fitting of a 3D point scatter and a CAD (face) models is also proposed. An example pf modular leats squares fitting two #D point scatters,based on simulated data, is known.
Keywords :
"Least squares methods","Computer Society","Computer vision","Pattern recognition"
Conference_Titel :
Computer Vision and Pattern Recognition - Workshops, 2005. CVPR Workshops. IEEE Computer Society Conference on
Print_ISBN :
0-7695-2372-2
DOI :
10.1109/CVPR.2005.487