PhD Thesis: Global Parametrization Algorithms for Quadmeshing
Quadrangular meshes are a central data structure in the domain of geometry processing, with applications ranging from computer graphics to numerical simulation. A promising approach for automatic generation of high quality quadmeshes relies on the observation that they consist in a deformation of the regular grid almost everywhere, except at a few singular vertices. Thanks to the computation of a parametrization, that is to say a flat representation, of the surface to be meshed, it is then possible to overlay a grid to be lifted back as a quad mesh. For quads to be extracted from this surface parametrization, it has to be seamless, that is to say satisfy a set of alignment constraints on its boundary and its cuts. This needs a quantization as integer values of some of its degrees of freedom. These constraints are usually enforced progressively in a well-studied pipeline of operations, consisting in the computation of a smooth frame field, defining the future singular vertices of the mesh, an integration phase from which a rotationally seamless parametrization is recovered, and a quantization step to determine the translational degrees of freedom.
This work was supervised by Dmitry Sokolov and Etienne Corman.
The defense occurred on December 6th 2023 at the Université de Lorraine.
Jury
- Xavier Goaoc, Université de Lorraine [President]
- Mirela Ben-Chen, Technion Israel Institute of Technology [Reviewer]
- Marco Tarini, University of Milan [Reviewer]
- Bruno Lévy, Inria, Centre de l’Université de Lorraine
- Pooran Memari, CNRS, Ecole Polytechnique
- Mélina Skouras, Inria, Centre de l’Université de Grenoble
- Etienne Corman, CNRS, LORIA [Advisor]
- Dmitry Sokolov, Université de Lorraine [Advisor]