Kolloquiumszyklus - previous dates

The human visual system (HVS) has its own limitations (e.g., the quality of eye optics, the luminance range that can be simultaneously perceived, and so on), which to certain extent reduce the requirements imposed on display devices. Still a significant deficit of reproducible contrast, brightness, spatial pixel resolution, and depth ranges can be observed, which fall short with respect to the HVS capabilities.  Moreover, unfortunate interactions between technological and biological aspects create new problems, which are unknown for real-world observation conditions.

In this talk, we are aiming at the exploitation of perceptual effects to enhance apparent image qualities.  At first, we show how the perceived image contrast and brightness  can be improved by exploiting the Cornsweet and glare illusions. Then, we present techniques for hold-type blur reduction, which is inherent for LCD displays.  Also, we investigate apparent resolution enhancements, which enable showing image details beyond the physical pixel resolution of the display device.  Finally, we discuss the problem of perceived depth enhancement in stereovision, as well as comfortable handling of specular effects, film grain, and video cuts.
 

We introduce a complete morphological analysis framework for 3D point clouds. Starting from an unorganized point set sampling a surface, we propose morphological operators in the form of projections, allowing to sample erosions, dilations, closings and openings of an object without any explicit mesh structure. Our framework supports structuring elements with arbitrary shape, accounts robustly for geometric and morphological sharp features, remains efficient at large scales and comes together with a specific adaptive sampler. Based on this meshless framework, we propose applications which benefit from the non-linear nature of morphological analysis and can be expressed as simple sequences of our operators, including medial axis sampling, hysteresis shape filtering and geometry-preserving topological simplification.

Computational Science and Visualization have been the focus of my studies and research for the last several years. The aim of Computational Science is to gain insight into complex systems and natural phenomena by capturing them in computational models and execute these models on the computer.  With Visualization I mean computer visualization in the broadest sense, including branches like Computer Graphics and Image Processing. During my talk I visit several of my projects, ranging from cloth simulation and fluid simulation to Deferred Shading and (non-photorealistic) rendering. One of the themes of my talk is how the programmable graphics processing unit narrowed the gap between realistic physically-based simulations and real-time computer graphics.

Streamlines and Streamsurfaces are standard tools for the visual analysis of flow data. Nevertheless, their applications still poses challenges concerning their extraction, integration, and visualization.

In the talk, we tackle three problems:
- the selection of suitable stream lines,
- a stable integration of stream surfaces,
- the selection of suitable stream surfaces.

We show that these problems can and should be formulated as global optimization problems. We present the respective error functionals to be minimized and show solutions for several test cases.
 

The vision of integrating the best of automated computation and capabilities of the human has been a highly praised goal in visualization research and parallels the emergence of visual analytics as a field on its own. In visual analytics, the integration of automated and interactive methods is considered to be one of the main mechanisms to facilitate the construction of knowledge in data analysis. This integration can be done at different levels -- from static visualizations of computation results to giving user the interactive control on the inner workings of an algorithm.

One form of such integrations involve the "seamless" use of automated computational tools within interactive visual data analysis. These methods employ "task-oriented" computations whose results become natural elements of the interactive process. Such approaches are important to build more reliable and insightful visual-data-analysis routines and to foster the use of visual analytics by a wider audience. In this talk, I will provide a quick overview of the different types of integration in visual analysis and focus on the examples of using computational tools seamlessly. The talk will then move on to discuss opportunities and open issues with such approaches.

The multidisciplinary BLUR project at UC Berkeley combines computer graphics with optics, optometry, and photography.  This research investigates mathematical models to describe the shape of the cornea and algorithms for cornea measurement, scientific and medical visualization for the display of cornea shape, mathematics and algorithms for the design and fabrication of contact lenses, simulation of vision using actual patient data measured by wavefront aberrometry, photo-realistic rendering algorithms for generating imagery with optically-correct depth of field, view camera simulation.  This talk will present an overview of rendering algorithms for simulating depth of field found in photographs and of vision-realistic rendering algorithms for simulating a subject's vision.   Recent work on correcting visual aberrations with computational light field displays will also be briefly introduced.

There are two parts to this talk. The first gives an overview of my Master’s thesis, which is about visualizing stream surfaces. The second is about modifications and additions to Paraview that I did as part of my internship at the Swiss National Supercomputing Center.

Stream surfaces are used to visualize snapshots in time of fluid flow. However, as useful as they are, they also have significant problems. In particular, humans have a difficult time understanding these surfaces and often make errors.

The thesis investigated the hypothesis that artificial coloring can help viewers perceive stream surfaces.  I present three algorithms for deciding on and placing colors based on a surface’s curvature or the use of simulated viewpoints. I will also discuss a user study that I designed and carried out that tested how users perceive images of surfaces that were created with these algorithms.

My work with Paraview involved improving the control of opacity in volume rendering. When using transfer functions to determine opacity in volume rendering, it can be difficult to get a good visualization with just one transfer function. Therefore, it can be useful to use two different transfer functions based on different aspects of the data, such as scalar and gradient values. I will describe and demonstrate a modification to Paraview’s GUI interface that allows users to use two different transfer functions and a new feature in Paraview that allows users to specify a two dimensional transfer function.

In the field of volume visualization the data is usually given as discrete samples. During rendering a continuous reconstruction is necessary. It is usually achieved by convolving the discrete data with a continuous filter. The assumption is that the discrete data is derived from a continuous signal sampled on a lattice, where the sinal is spherically bandlimited. In volume visualization Cubic Lattice (CC), the body-centered lattice (BCC) and the face-centered lattice (FCC) are used. For comparing the various filters on the lattices a benchmark signal, the famous ML-signal is commonly used. The authors arrived in the conclusion that the pre-aliasing effect is minimal on FCC lattice. However their assumption was based on the spherically band-limitedness of underlying signal. In our work we showed that the ML-signal is not spherically band limited, and it gives an unfair advantage to the FCC lattice during comparisons. On the other hand, we proposed that rotating the ML signal can lead to a fair comparison in terms of aliasing.

Interactive visualization and analysis of large and complex volume data is an ongoing challenge. Data acquisition tools produce hundreds of Gigabytes of data and are one step ahead of visualization and analysis tools. Therefore, the amount of data to be rendered is typically beyond the limits of current computer and graphics hardware performance. We tackle this challenge in the context of state-of-the-art out-of-core multiresolution direct volume rendering by using a common mathematical framework (a) to extract relevant features from these large datasets, (b) to reduce and compress the actual amount of data, and (c) to directly render the data from the framework coefficients. As a common framework, we introduced the higher-order extension of the matrix singular value decomposition - tensor approximation (TA) - as a compact volume data representation. In particular, the bases of tensor approximation were exploited to model state-of-the-art multiresolution volume visualization and multiscale feature extraction with one set of global bases. Based on this contribution, a feature scale metric was developed to automatically select a feature scale and a resolution for the final reconstruction. Thanks to the compact data representation by TA, a significant data compression and GPU-based real-time visualization was achieved. The new algorithms were tested on volume datasets from micro-computed tomography and phase-contrast synchrotron tomography that range up to 32 Gigabytes.