# Algebraic Relations among Matching Constraints of Multiple Images

1 MOVI - Modeling, localization, recognition and interpretation in computer vision
GRAVIR - IMAG - Graphisme, Vision et Robotique, Inria Grenoble - Rhône-Alpes, CNRS - Centre National de la Recherche Scientifique : FR71
Abstract : Given a set of $n\ge2$ uncalibrated views, for any corresponding point across $n$ views, there exist three types of matching constraints: bilinear constraints(for $n\ge2$), trilinear constraints (for $n\ge3$, \cite{shashua94c- }) and quadrilinear constraints (for $n \ge 4$, \cite{hartley94d,triggs95a,fau- geras95a}). The exact algebraic relations among these multi-linear constraints have not been elucidated by previous authors. This paper examines the relations between these matching constraints by singling out the degenerate view and point configurations. The key result that will be established is that for generic view configurations and generic points, all multi-linear constraints may algebraically be reduced to the algebraically independent bilinear constraints- . In other words, all matching constraints are contained in the ideal generated only by the bilinear constraints for generic views and points. As a consequence, $2n-3$ algebraically independent bilinearities from pairs of views completely describe the algebraic/geometric structure of $n$ uncalibrated views for generic views and points. For degenerate points of generic views, each type of constraint reduces differently. The exact reduced form of the matching constraints are also made explicit by computer algebra.
Keywords :
Type de document :
Rapport
RR-3345, INRIA. 1998
Domaine :

https://hal.inria.fr/inria-00073344
Contributeur : Rapport de Recherche Inria <>
Soumis le : mercredi 24 mai 2006 - 12:35:16
Dernière modification le : mercredi 11 avril 2018 - 01:52:10
Document(s) archivé(s) le : dimanche 4 avril 2010 - 23:43:11

### Identifiants

• HAL Id : inria-00073344, version 1

### Citation

Long Quan. Algebraic Relations among Matching Constraints of Multiple Images. RR-3345, INRIA. 1998. 〈inria-00073344〉

### Métriques

Consultations de la notice

## 133

Téléchargements de fichiers