Skip to Main content Skip to Navigation
Book sections

Multigrid convergence of discrete geometric estimators

Abstract : The analysis of digital shapes require tools to determine accurately their geometric characteristics. Their boundary is by essence discrete and is seen by continuous geometry as a jagged continuous curve, either straight or not derivable. {\em Discrete geometric estimators} are specific tools designed to determine geometric information on such curves. We present here global geometric estimators of area, length, moments, as well as local geometric estimators of tangent and curvature. We further study their {\em multigrid convergence}, a fundamental property which guarantees that the estimation tends toward the exact one as the sampling resolution gets finer and finer. Known theorems on multigrid convergence are summarized. A representative subsets of estimators have been implemented in a common framework (the library \texttt{DGtal}), and have been experimentally evaluated for several classes of shapes. The interested user has thus all the information for choosing the estimator best adapted to its application, and disposes readily of an efficient implementation.
Document type :
Book sections
Complete list of metadata
Contributor : Équipe Gestionnaire Des Publications Si Liris <>
Submitted on : Wednesday, August 10, 2016 - 4:16:32 PM
Last modification on : Thursday, March 25, 2021 - 3:18:08 AM


  • HAL Id : hal-01352952, version 1


David Coeurjolly, Jacques-Olivier Lachaud, Tristan Roussillon. Multigrid convergence of discrete geometric estimators. V. Brimkov & R. Barneva. Digital Geometry Algorithms. Theoretical Foundations and Applications to Computational Imaging, Springer, pp.395-424, 2012. ⟨hal-01352952⟩



Record views