Mise en correspondance par invariants locaux

Cordelia Schmid 1 Roger Mohr 1
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 : This paper presents an alternative to standard matching methods . Its main novel feature is to rely on greylevel invariants that are computed locally. This allows matching even in presence of large image change in rotation or scaling . These invariants are the set of differential invariants in rotation with an additional multiscale approach for taking into account scale changes . Such invariants were also expanded towards illumination changes allowing therefore large illumination variations . The overall method is quite easy to be understood and implemented and provides surprisingly good results .
Document type :
Conference papers
Complete list of metadatas

https://hal.inria.fr/inria-00548369
Contributor : Thoth Team <>
Submitted on : Monday, December 20, 2010 - 8:43:58 AM
Last modification on : Wednesday, April 11, 2018 - 1:53:37 AM

Identifiers

  • HAL Id : inria-00548369, version 1

Collections

IMAG | CNRS | INRIA | UGA

Citation

Cordelia Schmid, Roger Mohr. Mise en correspondance par invariants locaux. Reconnaissance des Formes et Intelligence Artificielle (RFIA '96), Jan 1996, Rennes, France. pp.236--245. ⟨inria-00548369⟩

Share

Metrics

Record views

109