Compression progressive et sans perte de structures géométriques

Gandoin Pierre-Marie 1
1 PRISME - Geometry, Algorithms and Robotics
CRISAM - Inria Sophia Antipolis - Méditerranée
Abstract : For a few years, meshes have been conquering a predominant status amongst the different types of computer data structure for geometric objects. More precisely, the meshes composed of simplices -- triangles for the modeling of surfaces embedded in 3D, tetrahedra for the modeling of volumes -- seem to be the most widespread at the present time. The rapid growth of applications using these geometric structures in such various fields as finite elements computing, simulation of surgery, or video games has quickly raised the problem of an efficient coding, well suited for storage or visualization. The advent of the World Wide Web, which needs a compact and progressive representation of the data in order to guarantee the user-friendliness of the exchanges and communications, increases this need. Therefore, since 1995, a large number of algorithms have been proposed for the compression of triangular meshes, using for most of them the following approach: the vertices of the mesh are coded in an order such that it contains partially the topology of the mesh. In the same time, some simple rules attempt to predict the position of the current vertex from the positions of its neighbors that have been previously coded. In this thesis, we have chosen to give the priority to the compression of the vertex positions rather than their connectivity. We describe a set of coding methods which are progressive, lossless, adapted to arbitrary meshes (non necessarily triangular nor manifold, with arbitrary genus), and generalizable to any dimension. Our results are competitive with regards to the best current progressive methods: for instance, for the particular case of triangular surface meshes, about 3.6 bits per vertex are necessary on average to code the connectivity of usual models.
Document type :
Theses
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-00771344
Contributor : Olivier Devillers <>
Submitted on : Tuesday, January 8, 2013 - 2:44:49 PM
Last modification on : Tuesday, February 26, 2019 - 11:19:49 AM
Long-term archiving on: Tuesday, April 9, 2013 - 3:53:03 AM

Identifiers

  • HAL Id : tel-00771344, version 1

Collections

Citation

Gandoin Pierre-Marie. Compression progressive et sans perte de structures géométriques. Géométrie algorithmique [cs.CG]. Université Nice Sophia Antipolis, 2001. Français. ⟨tel-00771344⟩

Share

Metrics

Record views

277

Files downloads

345