IMAGES GALERIE

Topological modelisation with multi-résolution (Pierre Kraemer)

Using MR-Maps in an multiresolution editing tool with subdivision surface using the Catmull-Clark scheme.

Using MR-Maps in an multiresolution editing tool with subdivision surface using the Loop scheme.

Using MR-Maps in an multiresolution editing tool with subdivision surface using an hybrid scheme.

Using MR-Maps in an multiresolution editing tool with subdivision surface using the Sqrt(3) scheme.

Meshing voxels (Dobrina Boltcheva)

Three key steps of the Discrete Delaunay reconstruction algorithm a) Discrete approximation of Voronoï area on the discrete border b) Discrete Voronoï graph c) Dual reconstruction : Euclidiean Voronï graph (in black), t)

Surface meshing of the "dragon" with three different resolutions.

Simultaneous meshing of the liver and the right kidney. The shared border is within the two meshes.

Example of meshing of a skeleton and aorta.

Detection and characterization of pockets in proteins (Benjamin Schwarz)

Given the structure of a protein, we search cavitys using a computational geometry. The geometric formalism allows to detect and manipulate easily the pockets. We determine, for exemple, the atoms participating in the vicinity of the pockets. The second part of the work consist in "meshing" proteins and their pockets using parametric patches to give a smooth surface representation.

The link domain of the receptor for vitamin D is represented with its edge representation, a translucent polygon highlight the main pocket.

Meshing scanned objects (Marc fournier)

Marching Cube new version and new Marching Square triangulation for vectorial distance transformation: 1.a) Standard Marching Cube triangulation (MC) with classical scalar distance. 1.b) same as 1.a) with vectorial distance. 1.c) building of the triangulation on the vectorial grid. 2.a) MC on scalar distance. 2.b) MC on vectorial distance. 3.a) Venus model. 3.b) MC on scalar distance. 3.c) MC on vectorial distance.

Meshing merging in the vectorial distance transformation model: 1.a) scanning the surface. 1.b) raw data with redondant measures. 1.c) triangulation results. 2.a) zoom on an area with redondant information. 2.b) is the result of the merging of the meshes.

Adaptive filtering for meshes using vectorial distance transformation: Adaptive filtering for meshes using vectorial distance transformation 1.a) initial data, 1.b) adding artificial noise 1.c) filtered data 2.a) initial data, including noise from scanner 2.b) filtered data.

Visualization of scanned objects (Frederic Larue)

Refraction is assimilated to a distortion depending on the point of view. This distortion is precomputed on the surface of the object. During the rendering, this precomputation is used to insure realtime visualisation.

Automatic methods have been proposed for acquisition. On the left a picture of the original model. In the middle, the geometry has been reconstructed. On the right : the render result.