Thomas Schiffer
 |
OFFICE
|
Motivation
Industry at large (like automotive or entertainment industry) demand photorealistic renderings of highly complex and dynamic scenes. These scene descriptions often contain computer vision-aquired data (like measured BRDFs or reconstructed geometry) for economic reasons and to increase the degree of realism. Generating high-quality images is a computationally intense task, which is made even more challenging by interactive or realtime time constraints.
Current algorithms (ray tracing, radiostity or photon mapping) do trace a huge amount of rays through the scene to compute a single image. Thus, a high performance ray casting engine serves as an essential building block of global illumination systems. The advent of increasingly parallel hardware raises the question how algorithms and data structures for ray casting can be mapped to these massively parallel architectures in an efficient and scalable manner.
Research
My PhD thesis is concerned with acquisition and photorealistic reproduction of dynamic scenes for interactive and/or real-time display.
My current research is focused on high performance ray casting on massively parallel hardware architectures (like NVidia’s CUDA technology). I am working on efficient building and traversal of acceleration structures for dynamic scenes, which can be subsequently used for stochastic ray tracing and radiosity calculations. My research results are implemented in the OpenGI software framework.
Teaching
Acknowledgements
My PhD position is funded by the FWF doctoral program 'Confluence of Vision and Graphics'.Publications
 |
T. Schiffer. A Parallel Geometry Core For High Performance Ray Tracing. Diploma Thesis, Technical University Graz, 2008
Raytracing is a widely used image synthesis technique with high computational costs. Modern graphics hardware not only offers enormous parallel computation power but also a fexible programming model. In this thesis CUDA, NVidia's new API for general purpose computations on graphics hardware, and its applications to ray tracing are discussed. A parallel ray tracing system based on CUDA is developed, which has a modular design and provides an extensible set of objects for scene modeling. The ray tracing system is based on a parallel geometry core, that exploits computation power of graphics hardware to perform geometric calculations. Test results show that the CUDA raytracer delivers almost interactive frame rates for scenes containing millions of geometric primitives. |
 |
F. Aurenhammer, M. Demuth, and T. Schiffer. Computing Convex Quadrangulations. In Proc. 5th Ann. Int. Symp. Voronoi Diagrams
in Science and Engineering, Voronoi's Impact on Modern Science, volume 4, pages 32-43, Kiev, Ukraine, 2008.
We use projected Delaunay tetrahedra and a maximum independent set approach to compute large subsets of convex quadrangulations on a given set of points in the plane. The new method improves over the popular pairing method based on triangulating the point set. |
