Kolloquiumszyklus - previous dates

A rapidly growing computer graphics community has contributed to dramatic increase in complexity with respect to geometry as well as physical phenomena. Simulating, approximating and visualizing geometry consisting of tens of millions of polygons simultaneously tested for collision or visibility is becoming increasingly common. Further, recent technological innovations from graphics card vendors have given impetus to achieving these results at very high frame rates. Despite tremendous developments in graphics hardware, capturing the complete surrounding environment poses a significant challenge. Given the added time constraint for real-time or interactive rates, simplified representations and suitable approximationsof physical effects are of key importance.
This talk will focus on simplified representations and computations to achieve real-time performance for complex tasks and concentrates on a variety of topics including simplification, visibility, soft shadows and voxelization.

The talk will be about the physiology of plants and how to implement branching, tropisms, hormons and their control in simplifyfied simulations. Also, generation for real time models including metric, LODs, textures, streaming as well as rendering and lighting will be covered in the talk.

Numerical flow visualization is gaining importance because of the continuing trend from experiments toward computational fluid dynamics. We have come to the situation where reliable numerical data are easily available but often hard to interpret because their size and intricacy challenge current visualization tools. The state of the art in flow visualization is advancing on several fronts, an important one being the field of feature-based visualization, which aims at revealing flow features such as vortices, flow separation, or recirculation. Such flow phenomena are of interest because of their effect, either beneficial or adverse, in industrial applications like power generation, mixing, or combustion. Feature-based flow visualization again splits into several branches, but one of them has become particularly popular under the name of vector field topology.

In this talk we present work in topology-based flow visualization, resulting from our collaboration with turbomachinery companies and focusing on the optimization of water turbines. We discuss the usage of vector field topology for extracting the above mentioned flow features, and additional ones such as vortex rings and vortex breakdown bubbles. We address the limitations of vector field topology and the current search for an adequate extension to unsteady flow fields. Finally we move to the field of Lagrangian coherent structures, which can be interpreted as a time-dependent variant of vector field topology. There, we present a technique for accelerating their computation, based on adaptive mesh refinement.

Few scientific topics catch as much attention as climate research these days. Will temperatures rise significantly in the upcoming decades? Will snow and ice covers disappear? Will draugths and severe storms threaten lifes all around the globe? Many questions like these move people, politics, and also business. To give as good as possible answers, climate researchers employ most modern measurement, simulation, and analysis methodology, resulting in challenging compilations of data of heterogeneous form and origin, usually with multi- ple variates, and almost always time-dependent. Accordingly, visualization is challenged and advanced approaches are needed to enable effective exploration, analysis and presen- tation. In this talk recent research work on how to support hypothesis generation in climate research through interactive visual exploration is presented. A discussion of associated challenges explains why visualization of data from climate research clearly has the potential to initiate interesting future research in visualization.

The video game industry is the only entertainment industry that has seen steady, double digit growth over the past fifteen years even out doing the high tech industry.
This provides many opportunities for talented individuals who may not have thought about making games as a career.
Wolfgang Hamann (Pres/CEO - Koolhaus Games, Vancouver, Canada) will be providing an overview of this fascinating field as well as discussing the Game Development Process - the marriage of software development, game design, art and sound

3D datasets are becoming increasingly large and complex. In medicine, fiber structures within the brain are inferred from diffusion-tensor magnetic resonance imaging (DT-MRI), yielding thousands to millions of curved trajectories. In chemistry, billions of atoms are included in large-scale molecular dynamics simulations. In both cases, the resulting geometry becomes difficult to comprehend in part because of its complexity.
We describe two approaches to improving perception of the resulting 3D scenes. The first approach is to apply physically based illumination rather than the conventional "local" illumination. The second approach is to transform the data into an ensemble coordinate system where geometric complexity increases slowly. The utility of these approaches have been validated by performing user studies.

In 2005 the university hospital of Maastricht (Netherlands) acquired a Medtronic N20 Polestar, a mobile, low-field, intraoperative MRI scanner which can be used in a standard operating room. To make better use of this scanner and enhance its images with high-quality preoperative imaging data the hospital started a close collaboration with the biomedical image analysis group of the technical university of Eindhoven (prof. Bart ter Haar Romeny). The goal is to define research projects related to image-guided neurosurgery that involve preoperative imaging, intraoperative MRI or a combination of both. For this purpose, the image-guided neurosurgery group was founded in 2007. This group is directly located at the university hospital of Maastricht and therefore at only a minute's distance from the neurosurgeons, one of whom (a resident) is a part-time but active member of this group working on his own PhD. This KV presentation will discuss the image-guided neurosurgery group and some of its past and current projects. Special attention will be given to deep brain stimulation and multimodal visualization, the latter of which was a master's thesis project of the presenter Ralph Brecheisen. His own PhD project in the image-guided neurosurgery group has only recently started and is still ill-defined but will focus on visualization for neurosurgery applications. The main purpose of his visit to Vienna is to find ideas for research and opportunities for collaboration, possibly in combination with the university hospital of Maastricht.

High-fidelity real-time visualization of surfaces under high-dynamic-range (HDR) image-based illumination provides an invaluable resource for various computer graphics applications. Material design, lighting design, architectural previsualization, and gaming are just a few such applications.
We present filtered importance sampling, a technique for image-based lighting of glossy objects using BRDF importance sampling in conjunction with environment map filtering. Furthermore, we extend the algorithm with real-time shadow computation. Free from any pre-computation, the algorithm supports fully dynamic scenes and, above all, is simple to implement.

The goal of computerized image generation is to convey information to the user or viewer, whether for artistic reflection, scientific discovery, or decision making. The history of art, design, illustration, and perception form a rich basis for developing interactive computerized visual environments for discovery, engineering, and analytical decision making. The development of interactive visualization techniques to effectively, rapidly and accurately convey information can fundamentally change the pace of scientific discovery and adoption of new science into usable applications. Moreover, integrated data management, analysis, and interactive visual environments provide insight and information from the massive data generated by computer simulations, sensors, and web-based sources. The potential of these integrated environments has led to a evolution of visualization to visual analytics.
In this talk, I'll discuss the role that computer graphics and interactive integrated visualization and analytics can play in research, discovery, and deployment in a variety of application areas. We have been developing integrated visual analytics environments for a diverse set of applications from homeland security to fundamental computational science. I'll describe some of our integrated visualization, data management, and analytical solutions for weather forecasting, cloud modeling, surgical training, computational nanoelectronics, computational fluid dynamics, cancer care engineering, syndromic surveillance, agricultural food production, and emergency response.

Point counting is a technique used for estimate the composition of a sample. We developed an interactive application designed to facilitate the point counting process through visual computing; this application could be useful in any domain in which the sample image can be digitized. Additionally, we provide other visual analysis tools showing how the application can benefit from the combination of these concepts and the techniques coming from Information Visualization. We utilize the point counting technique on petrographic thin section images in order to perform modal analyses of rocks, but it could be used on many other types of samples in order to estimate their composition.

With the recent advancements in biomedical imaging technology, molecular processes in the cell can now be brought directly in relation to structural and functional changes at higher levels. Functional imaging (optical, nuclear and MR with targeted contrast agents) provides a window on cellular biochemistry and gene expression, while structural imaging (CT, MR, ultrasound) may be used to measure the resulting structural changes in the whole body. Disease processes and treatment effects can now be followed over time, from molecule to organism, both in pre-clinical small animal models and in humans.
For a single time point, a molecular imaging study may consist of photographs, photon emission images, serial CT, MR or PET slices, functional MR imaging, MR spectroscopy, and histology. We have recently started working on new visualization techniques that will enable the in-depth visual exploration of relations between disease evolution, underlying molecular processes and structural and functional changes that are locked up in combined molecular, functional and structural imaging data.
This talk starts with a brief overview of the TU Delft Medical Visualisation group and its research activities. I will then introduce molecular imaging and the role of visualisation in this new field, after which I will present our latest results on a visualisation system for bioluminescence imaging and on whole-body articulated registration for small animal imaging.

In this talk we show how biased methods based on density estimation of photon hits such as photon maps can be extended. Instead of using photon hits as in photon maps and combining them with final gathering we exploit the photon paths in two ways. First, we make density estimation in ray space of photon paths, which reduces bias. Second, we reverse the process of density estimation and splat energy of photons around the cones of photon paths. We show the results and timings of the image synthesis based on the two new methods and the differences to traditional photon mapping.

The appearance of real-world objects depends on the incident illumination, on the 3D geometry of the object, and on the reflection properties of the object's surfaces. Reflectance fields capture the resulting global light transport in such a way that the object can be relit in arbitrary virtual environments faithfully reproducing the appearance of the original. In this talk I will present an overview about our current work on acquiring and processing reflectance fields. One part will cover acquisition techniques that are able to measure the global light transport within a scene on a ray-to-ray basis allowing for capturing and reproducing effects such as subsurface scattering, refractions and caustics. One remaining problem is that reflectance fields are typically acquired only for a discrete set of incident light directions. A rotation in the incident illumination is likely to produce artifacts due to this coarse sampling. In the second part I therefore will address the problem of upsampling reflectance fields in the light domain allowing for smoothly moving shadows and highlights when light sources move around the scene.