Matching Perspective Images Using Geometric Constraints and Perceptual Grouping

Abstract : This paper presents a method for matching two perspective images from a well defined indoor scene. The method insists on exploring both geometric constraints and perceptual grouping for reducing the search space. The geometric constraints used are principally perspective information, such as vanishing point, horizon line and projective coordinates. The perceptual groups are such as directional, rays and collinear groups. Our method succeeds in analyzing successively the rotational and translational effect. which makes geometric constraints more direct and easier to explore. The matching strategy is coarse-to-fine, based on the hierarchical perceptual grouping. The hypotheses are then propagated and verified to the whole image by neighboring, relationship without backtracking. The matching results are also presented.
Type de document :
Communication dans un congrès
2nd International Conference on Computer Vision (ICCV '88), Dec 1988, Tampa, United States. IEEE Computer Society, pp.679--683, 1988, 〈http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=590050〉. 〈10.1109/CCV.1988.590050〉
Liste complète des métadonnées

https://hal.inria.fr/inria-00548476
Contributeur : Thoth Team <>
Soumis le : lundi 20 décembre 2010 - 08:48:36
Dernière modification le : jeudi 11 janvier 2018 - 06:23:18

Identifiants

Collections

Citation

Long Quan, Roger Mohr. Matching Perspective Images Using Geometric Constraints and Perceptual Grouping. 2nd International Conference on Computer Vision (ICCV '88), Dec 1988, Tampa, United States. IEEE Computer Society, pp.679--683, 1988, 〈http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=590050〉. 〈10.1109/CCV.1988.590050〉. 〈inria-00548476〉

Partager

Métriques

Consultations de la notice

51