David Cazier
Home ||
Research Works ||
Projects and Collaborations ||
Publications ||
Curriculum Vitae
Crowds Simulation ||
Collision Detection ||
Multiresolution models ||
Mesh Reconstruction ||
Non manifold models
Assistant Professor in Computer Science
- Reseacher at the LSIIT laboratory
- Teacher at IUT de Haguenau and at Université de Strasbourg.
- Follow me on Facebook or LinkedIn
Phone LSIIT : | +33 (0)3-68-85-45-68 | Phone IUT : | +33 (0)3-88-05-34-31 |
Fax LSIIT : | +33 (0)3-68-85-44-55 | eMail : | david.cazier AT unistra.fr |
Geometric models for the Virtual Reality: Representation // Simulation // Interaction
Source of innovation, subject of futuristic scenarios, bearing of dreams, virtual reality fascinates the general public and feeds the research activities of many laboratories around the world. Based on the immersion of users and the notions of interactivity and real-time, the virtual reality makes it possible to to immerse oneself in a synthesis environment to better understand, build or modify reality. It is used in many sectors: from architecture to urban planning, to health, research and industry.
Create realistic virtual worlds with which users can interact in real time requires the development of geometric models of increasing complexity, supporting a wide range of simulations (behavioral, mechanical, multi-physics). Often to speed up numerical processing, rendering and interaction possibilities, they are associated with hierarchical or multi-scale structures.
My research took place in this framework and are divided into three axes:
- generic and multiresolution models adapted to a wide range of meshes;
- structuring of space to improve real-time interaction (combinatorial topology);
- development of tools to generate such models and adapt them to specific treatments.
Activités de Recherche
Many works in simulation, modeling or geometry processing use multiple representations of the same object. They may correspond to different scales or to different levels of detail, especially in multiresolution applications. They may also correspond to models of different natures.
In simulation, most methods require volumetric meshing for the computation of deformation. To this first mesh a finer surface mesh is usually associated for realistic rendering. The use of multi-scale representation is also common in the field of image segmentation for mesh compression or simplification. At last, many algorithms make use of hierarchical structures to speed up the treatments, such as ray tracing and collision detection.
My research concerns the development of multi-scale combinatorial structures for geometric modeling, animation and geometry processing.
A compact and efficient combinatorial model for the encoding of multiresolution surfaces. This work is applied to subdivision surfaces and progessive meshes.
Algorithms for the automatic building of manifold and volumetric meshes from medical images.
A new model for the modeling of curves, surfaces and volumes sewn and assembled around singular points (also called non-manifolds models)
A forecast mechanism for real time tracking of particles trajectories. This work is applied to collision detection between deformable solids in complex environments.