This paper presents a new framework for solving geometric structure and motion problems based on L/sub /spl infin//-norm. Instead of using the common sum-of-squares cost-function, that is, the L/sub /spl infin//-norm, the model-fitting errors are measured using the L/sub /spl infin//-norm. Unlike traditional methods based on L/sub 2/ our framework allows for efficient computation of global estimates. We show that a variety of structure and motion problems, for example, triangulation, camera resectioning and homography estimation can be recast as a quasiconvex optimization problem within this framework. These problems can be efficiently solved using second order cone programming (SOCP) which is a standard technique in convex optimization. The proposed solutions have been validated on real data in different settings with small and large dimensions and with excellent performance.
Multiple view geometry and the L/sub /spl infin//-norm
Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1 ; 2 ; 1002-1009 Vol. 2
2005-01-01
453192 byte
Conference paper
Electronic Resource
English
Multiple View Geometry and the L-norm
British Library Conference Proceedings | 2005
|Linear two-view multiple plane geometry
British Library Online Contents | 2008
|Generalized Convexity in Multiple View Geometry
British Library Online Contents | 2010
|H/sub /spl infin// control of active vehicle suspensions
IEEE | 2000
|