Geometric and topological shape description using discrete curvature
Event details
| Date | 11.02.2020 |
| Speaker | Jean-Luc Mari |
| Location | |
| Category | Conferences - Seminars |
Abstract: We present two approaches to extract geometric features on meshes, based on the use of discrete curvature estimators in order to define specific areas of interest before processing these meshes. The first part deals with the computation of feature lines using a homotopic thinning algorithm transposed on specific regions of a mesh. To do this, the notion of neighborhood is transposed from classical thinning algorithms (where the adjacency is constant) to the mesh domain where the neighborhood is variable due to the adjacency of each vertex. In the second part, we present a new shape descriptor for 3D meshes called SCG for ”Shape-Curvature-Graph”. It aims at representing an arbitrary triangular polyhedron using a graph. A discrete curvature map is computed and an adjacency graph is constructed with a node for each patch. We show that this specific shape-graph can be used to perform self-similarity detection, or more generally to extract patterns within a shape.
Speaker: Jean-Luc Mari, Aix-Marseille University
Information about time and place of the Applied Topology Seminar can be found on the webpage: https://www.epfl.ch/labs/hessbellwald-lab/seminar/apptopsem1920/
Speaker: Jean-Luc Mari, Aix-Marseille University
Information about time and place of the Applied Topology Seminar can be found on the webpage: https://www.epfl.ch/labs/hessbellwald-lab/seminar/apptopsem1920/
Links
Practical information
- Informed public
- Free
Organizer
- Applied Topology Seminar