Skip to Main content Skip to Navigation

Adaptation de schémas de subdivision pour la reconstruction d'objet sans artefact

Abstract : In this thesis we analyse the behaviour of subdivision schemes, tool used to modelise smooth and multi-resolution surfaces. Firstly, our work aimed to study the geometric behaviour of these surfaces on the neighbourhood of the extraordinary vertices in the control mesh. The geometry of a subdivision surface is complex on the neighborhood of an extraordinary vertex, some of the unpleasant behaviours are obscure. We have proposed an evaluation framework for subdivision surfaces throughout an suitable measure : the absolute curvature gradient. Then we have proposed an interrogation span suitable to the analysis of an extraordinary vertex neighborhood. It is independent of the subdivision scheme used to synthesize the surface, thus we can compared several schemes. Then our work engaged in the polar spectral analysis of these geometric behaviours, taking into account there radial and angular characteristics regarding the topology of an extraordinary vertex neighbourhood. Our analysis extends the existing works for the interrogation of this modelisation tool. Secondly, we have proposed a description framework for the topological step of a subdivision scheme. Our system takes the form of a compact and flexible coding, it generalizes the existent descriptions. This coding allows the description of the topological step of all the known subdivision schemes, and many others.
Complete list of metadatas
Contributor : François Destelle <>
Submitted on : Wednesday, November 10, 2010 - 10:41:50 AM
Last modification on : Thursday, November 19, 2020 - 12:59:39 PM
Long-term archiving on: : Friday, February 11, 2011 - 2:48:19 AM


  • HAL Id : tel-00534612, version 1


François Destelle. Adaptation de schémas de subdivision pour la reconstruction d'objet sans artefact. Modélisation et simulation. Institut National Polytechnique de Grenoble - INPG, 2010. Français. ⟨tel-00534612⟩



Record views


Files downloads