de.mpg.escidoc.pubman.appbase.FacesBean
English
 
Help Guide Disclaimer Contact us Login
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT

Released

Conference Paper

A Point-based Approach to PDE-based Surface Reconstruction

MPS-Authors
http://pubman.mpdl.mpg.de/cone/persons/resource/persons44929

Linz,  Christian
Computer Graphics, MPI for Informatics, Max Planck Society;
Graphics - Optics - Vision, MPI for Informatics, Max Planck Society;

http://pubman.mpdl.mpg.de/cone/persons/resource/persons44508

Goldluecke,  Bastian
International Max Planck Research School, MPI for Informatics, Max Planck Society;
Graphics - Optics - Vision, MPI for Informatics, Max Planck Society;

http://pubman.mpdl.mpg.de/cone/persons/resource/persons44965

Magnor,  Marcus
Graphics - Optics - Vision, MPI for Informatics, Max Planck Society;

Locator
There are no locators available
Fulltext (public)
There are no public fulltexts available
Supplementary Material (public)
There is no public supplementary material available
Citation

Linz, C., Goldluecke, B., & Magnor, M. (2006). A Point-based Approach to PDE-based Surface Reconstruction. In K. Franke, K. R. Müller, B. Nickolay, & R. Schäfer (Eds.), Pattern Recognition : 28th DAGM Symposium (DAGM'06) (pp. 729-738). Berlin, Germany: Springer.


Cite as: http://hdl.handle.net/11858/00-001M-0000-000F-24A7-5
Abstract
Variational techniques are a popular approach for reconstructing the surface of an object. In previous work, the surface is represented either implicitly by the use of level sets or explicitly as a triangle mesh. In this paper we describe new formulations and develop fast algorithms for surface reconstruction based on partial differential equations (PDEs) derived from variational calculus using an explicit, purely point-based surface representation. The method is based on a Moving Least-Squares surface approximation of the sample points. Our new approach automatically copes with complicated topology and deformations, without the need for explicit treatment. In contrast to level sets, it requires no postprocessing, easily adapts to varying spatial resolutions and is invariant under rigid body motion. We demonstrate the versatility of our method using several synthetic data sets and show how our technique can be used to reconstruct object surfaces from real-world multi-view footage.