Projects |  Graphics |  Art |  Linux Helpers |  Links
Webmaster: 
© Nicolas Cuntz

Projects Graphics Art Linux Helpers Links
Interactive Time-Adaptive Flow Lines:
images videos


demo video
52 MB, 640 x 480, 4000.4 kbps
(codec: Xvid, DivX compatible)

Stream, path and streak lines are known to be very useful for the visualization of unsteady flows. Despite their importance in practice, appropriate algorithms suited for contemporary hardware are rare. In particular, the adaptive construction of the different line types is not sufficiently studied.

This work provides a profound representation and discussion of stream, path and streak lines. Two algorithms are proposed for efficiently and accurately generating these lines using modern graphics hardware. Each includes a scheme for adaptive time-stepping. The adaptivity for stream and path lines is achieved through a new processing idea we call selective transform feedback. The adaptivity for streak lines combines adaptive time-stepping and a geometric refinement of the curve itself. The approach is evaluated by applying it to analytically defined and texture-based examples.

Dynamic Flow Volumes using Particle Level Sets:
images videos


demo video
9 MB, 1024 x 768, 1456.3 kbps
(codec: Xvid, DivX compatible)

Typically, flow volumes are represented and visualized by describing their boundary as the iso-surface of a level set function. The particle level set (PLS) method combines the advantages of both grid- and particle-based level set representations.

This work demonstrates that the PLS method can be adapted to volumetric dye advection via streak volumes, and to the visualization by time surfaces and path volumes. This is achieved with a modified and extended PLS, including a model for dye injection. A new algorithmic interpretation of PLS is introduced in order to exploit the efficiency of modern graphics hardware, leading to an interactive visualization. Finally, the high quality and usefulness of PLS flow visualization is demonstrated by providing quantitative results on volume preservation and by discussing typical applications of 3D flow visualization.

GPU-based Particle Flow:
images videos

more images: VR flow-vis

demo video
101 MB, 640 x 480, 3993.6 kbps
(codec: Xvid, DivX compatible)

Direct flow visualizations can help climate researchers to explore flow characteristics by adjusting parameters and thus controlling the visual output. They provide an intuitive insight into complex flow data sets such as simulated circulations of the atmosphere or the ocean. Recent developments in visualization techniques exploiting programmable features of current graphics processing units (GPUs) have proven to be very powerful. This is especially true for particle-based techniques.

This work presents a complete GPU-based particle engine and framework for the real-time visualization of unsteady climate flow data sets. This involves a proper data work flow from the simulation back-end to the visualizer, the handling of non-uniform data grids and, finally, a proper support for the interactive exploration. We evaluate our framework with a South Asian typhoon simulated by the DKRZ Hamburg.

The application has been extended to multi-node rendering in order to immerse the user into the virtual reality cave at the University of Siegen. The cave consists of a cylindrical projection area composed of six (stereo) projectors (four front, two floor). An enhanced interactivity is achieved by introducing head and object tracking to the user interface.

GPU-based Particle Level Sets:
images videos

more images: Zalesak's sphere, VR clay bunny

demo video
25 MB, 640 x 480, 2472.5 kbps
(codec: Xvid, DivX compatible)

Level sets are used for the representation and evolution of closed surfaces. Grid-based level sets offer a good global representation which deals well with topological changes, but they suffer from numerical diffusion, whereas particle-based methods preserve details more accurately but introduce the problem of unequal global representation. The particle level set (PLS) method combines the advantages of both approaches by interchanging the information between the grid and the particles.

This work presents an enhanced PLS approach which fully maps to the GPU. Improvements w.r.t. the original PLS technique include a sub-voxel interface representation and an accurate level set correction using more precise particle radii. Compared to a public CPU-based reference implementation, our method achieves both, higher performance and superior quality in terms of volume preservation.

As a concrete application we demonstrate that our fast and accurate PLS is also well-suited for the visualization of dynamic flows.

Dynamic Particle Coupling:
images videos

demo video
35 MB, 640 x 480, 4016.2 kbps
(codec: Xvid, DivX compatible)

The main research approaches in computational fluid dynamics (CFD) are grid-based or particle-based. For interactive fluid simulations, grid-based techniques for the use of Graphics Processing Units (GPUs) have been developed in order to speed up the simulation.

This paper describes an approach for setting up a particle-based fluid simulation on the GPU. This builds upon earlier results on simulation of uncoupled particles. The major contribution of this work is a new approach for modeling dynamic particle coupling solely based on individual particle contributions. This technique does not need global sorting or an explicit solution of the n-nearest neighbor problem.


Publications:
2009:
Nicolas Cuntz, Albert Pritzkau, Andreas Kolb:
Time-Adaptive Lines for the Interactive Visualization of Unsteady Flow Data Sets,
in: Computer Graphics Forum (accepted for publication), 2009

Iurie Chiosa, Andreas Kolb, Nicolas Cuntz, Marvin Lindner:
Parallel Mesh Clustering,
in: Proc. Eurographics Symp. on Parallel Graphics and Visualization, p. 33-40, 2009

2008:
Nicolas Cuntz, Andreas Kolb, Robert Strzodka, Daniel Weiskopf:
Particle Level Set Advection for the Interactive Visualization of Unsteady 3D Flow (pdf),
in: Computer Graphics Forum, 27 (3), p. 719-726, 2008

2007:
Nicolas Cuntz, Martin Leidl, Andreas Kolb, Christof Rezk-Salama, Michael Böttinger:
GPU-based Dynamic Flow Visualization for Climate Research Applications (pdf),
in: Proc. Simulation and Visualization , 2007

Nicolas Cuntz and Andreas Kolb:
Fast Hierarchical 3D Distance Transforms on the GPU (pdf),
in: Eurographics (Short Paper), 2007

2005:
Andreas Kolb and Nicolas Cuntz:
Dynamic Particle Coupling for GPU-based Fluid Simulation,
in: Proc. 18th Symp. on Simulation Technique, p. 722-727, 2005

Posters:
2007:
Fast Hierarchical 3D Distance Transforms on the GPU (pdf),
presented at the Eurographics 07 Poster Session, 2007

Parallel Particle Level Set Method on the GPU (pdf),
I3D 07 Poster Session, 2007