Scientific Visualization

The goal of visualization is to effectively and accurately communicate data. Visualization research has often overlooked the errors and uncertainty which accompany the scientific process and describe key characteristics used to fully understand the data. The lack of these representations can be attributed, in part, to the inherent difficulty in defining, characterizing, and controlling this uncertainty, and in part, to the difficulty in including additional visual metaphors in a well designed, potent display. However, the exclusion of this information cripples the use of visualization as a decision making tool due to the fact that the display is no longer a true representation of the data. This systematic omission of uncertainty commands fundamental research within the visualization community to address, integrate, and expect uncertainty information. In this chapter, we outline sources and models of uncertainty, give an overview of the state-of-the-art, provide general guidelines, outline small exemplary applications, and finally, discuss open problems in uncertainty visualization.

Color vision deficiency (CVD) affects a large number of individuals around the world, compromising their ability to effectively interpret color-coded information. This directly impacts the way these individuals perceive visualizations, often introducing ambiguities and uncertainties. This article provides an overview of the causes of color vision deficiency and discusses the main tools and techniques available for helping designers to create more effective visualizations for individuals with CVD. It also discusses the limitations of the existing techniques and presents some open questions for guiding research efforts in improving visualization experiences for larger audiences.

One of the greatest challenges for today’s visualization and analysis communities is the massive amounts of data generated from state of the art simulations. Traditionally, the increase in spatial resolution has driven most of the data explosion, but more recently ensembles of simulations with multiple results per data point and stochastic simulations storing individual probability distributions are increasingly common. This chapter describes a relatively new data representation for scalar data, called hixels, that stores a histogram of values for each sample point of a domain. The histograms may be created by spatial down-sampling, binning ensemble values, or polling values from a given distribution. In this manner, hixels form a compact yet information rich approximation of large scale data. In essence, hixels trade off data size and complexity for scalar-value “uncertainty”.

We discuss strategies to successfully work with strict feature definitions such as topology in the presence of noisy/uncertain data. To that end, we review previous work from the literature and identify three strategies: the development of fuzzy analogs to strict feature definitions, the aggregation of features, and the filtering of features. Regarding the latter, we will present a detailed discussion of filtering ridges/valleys and topological structures.

In scientific visualization the key task of research is the provision of insight into a problem. Finding the solution to a problem may be seen as finding a path through some rugged terrain which contains mountains, chasms, swamps, and few flatlands. This path—an algorithm discovered by the researcher—helps users to easily move around this unknown area. If this way is a wide road paved with stones it will be used for a long time by many travelers. However, a narrow footpath leading through deep forests and deadly swamps will attract only a few adventure seekers. There are many different paths with different levels of comfort, length, and stability, which are uncertain during the research process. Finding a systematic way to deal with this uncertainty can greatly assist the search for a safe path which is in our case the development of a suitable visualization algorithm for a specific problem. In this work we will analyze the sources of uncertainty in heuristically solving visualization problems and will propose directions to handle these uncertainties.

Predictive models are used in many fields to characterize relationships between the attributes of an instance and its classification. While these models can provide valuable support to decision-making, they can be challenging to understand and evaluate. While they provide predicted classifications, they do not generally include indications of confidence in those predictions. Typical quality measures for predictive models are the percentage of predictions which are made correctly. These measures can give some insight into how often the model is correct, but provide little help in understanding under what conditions the model performs well (or poorly). We present a framework for improving understanding of predictive models based on the methods of both machine learning and data visualization. We demonstrate this framework on models that use attributes about individuals in a census data set to predict other attributes of those individuals.

Network security defense often involves uncertain data which can lead to uncertain judgments regarding the existence and extent of attacks. However, analytic uncertainty and false positive decisions can be integrated into analysis tools to facilitate the process of decision making. This paper presents an interactive method to specify and visualize uncertain decisions to assist in the detection process of network intrusions. Uncertain decisions on the degree of suspicious activity for both temporal durations and individual nodes are integrated into the analysis process to aide in revealing hidden attack patterns. Our approach has been implemented in an existing security visualization system, which is used as the baseline for comparing the effects of newly added uncertainty visualization component. The case studies and comparison results demonstrate that uncertainty visualization can significantly improve the decision making process for attack detection.

Fiber tracking based on diffusion weighted Magnetic Resonance Imaging (dMRI) allows for noninvasive reconstruction of fiber bundles in the human brain. In this chapter, we discuss sources of error and uncertainty in this technique, and review strategies that afford a more reliable interpretation of the results. This includes methods for computing and rendering probabilistic tractograms, which estimate precision in the face of measurement noise and artifacts. However, we also address aspects that have received less attention so far, such as model selection, partial voluming, and the impact of parameters, both in preprocessing and in fiber tracking itself. We conclude by giving impulses for future research.

Uncertain field visualization is currently a hot topic as can be seen by the overview in this book. This article discusses a mathematical foundation for this research. To this purpose, we define uncertain fields as stochastic processes. Since uncertain field data is usually given in the form of value distributions on a finite set of positions in the domain, we show for the popular case of Gaussian distributions that the usual interpolation functions in visualization lead to Gaussian processes in a natural way. It is our intention that these remarks stimulate visualization research by providing a solid mathematical foundation for the modeling of uncertainty.

A challenge, visualization is often faced with, is the complex structure of scientific data. Complexity can arise in various ways, from high dimensionalities of domains and ranges, time series of measurements, ensemble simulations, to heterogeneous collections of data, such as combinations of measured and simulated data. Many of these complexities can be subsumed under a concept of multifields, and in fact, multifield visualization has been identified as one of the major current challenges in scientific visualization. In this chapter, we propose a multifield definition, which will allow us a systematic approach to discussing related research.

Multifield visualization covers a range of data types that can be visualized with many different techniques. We summarize both the data types and the categories of techniques, and lay out the reasoning for dividing this Part into chapters by technique rather than by data type.

In this chapter, we consider the need in multifield visualization to depict information contained in two or more fields in a compositional manner. There are many different visual channels, some of which are more commonly seen in visualization than others. Channel fusion occurs when two or more visual entities have to share the same screen space. By applying appropriate constructive operations on visual channels in the composition, one may encode the integration as well as separation of the underlying information depicted by the original channels. One special situation is where multiple fields are a set of fields from different temporal steps, which imposes additional constraints on the use of visual channels. It is inevitable that the availability of visual channels will not be able to scale up to a large number of visual channels. Hence, we consider briefly several general-purpose data mapping methods that can be used to reduce the complexity of visual mapping.

In this chapter, we present a state of the art on glyph-based visualization techniques that address the complex challenges of multi-field visualization. Glyphs are discrete parametrized visualization objects that encode multiple data values based on appearance (i.e., visual channels) such as size, shape, color, and opacity, and are effective for conveying multiple fields of data simultaneously. We provide a categorization of these techniques with the aim for an informative overview of recent literature. Our categorization is based on visual channels utilized by the glyph for mapping each data attribute, and the spatial dimensionality of the glyph-based visualization. We also discuss critical design aspects of glyph-based visualization to deal with the perceptual challenges inherent with this approach.

This chapter reviews various methods for multifield visualization that are based on the notion of derived fields. The derived fields are categorized based on properties like the number and type of input fields. Mathematical properties, algorithms, and applications are discussed for each derived field. Correlation and alignment measures are examined for a set of homogeneous fields, including pairwise similarity/dissimilarity measurements. Multifield analysis is also discussed in the context of input fields being the components of the decomposition of another field, possibly of a different type. Finally, research challenges are discussed in the context of the design of multifield analysis and visualization methods based on the concept of derived fields.

Interactive exploration and analysis of multi-field data utilizes a tight feedback loop of computation/visualization and user interaction to facilitate knowledge discovery in complex datasets. It does so by providing both overview visualizations, as well as support for focusing on features utilizing iterative drill-down operations. When exploring multi-field data, interactive exploration and analysis relies on a combination of the following concepts: (i)

physical views

that show information in the context of the spatiotemporal domain (domain perspective), (ii)

range views

show relationships between multiple fields (range perspective), and (iii) selecting/marking data subsets in one view (e.g., regions in a physical view) leading to a consistent highlighting of this subset in all other views (brushing and linking). Based on these principles, interactive exploration and analysis supports building complex feature definitions, e.g., using Boolean operations to combine multiple selections. Utilizing derived fields, statistical methods, etc., adds a further layer of flexibility to this approach. Using these concepts, it is also possible to integrate feature detection methods from the other chapters of this part, as well as application-specific feature extraction methods into an joint framework. This methodology of interactive visual data exploration and analysis has proven its potential in a larger number of successful applications. It has been implemented in a larger number of systems and is already available for a wide spectrum of different application domains.

The attribute space of a multivariate volume data set is hard to handle interactively in the context of volume visualization when more than three attributes are involved. Automatic or semi-automatic approaches such as involving clustering help to reduce the complexity of the problem. Clustering methods segment the attribute space, and the segmentation can be exploited for visual exploration of the volume data. We discuss user-guided and automatic clustering approaches of the multi-dimensional attribute space and visual representations of the results. Coordinated views of object-space volume visualization with attribute-space clustering results can be applied for interactive visual exploration of the multivariate volume data and even for interactive modification of the clustering results. Respective methods are presented and discussed and future directions are outlined.

Feature-based techniques are one of the main categories of methods used in scientific visualization. Features are structures in a dataset that are meaningful within the scientific or engineering context of the dataset. Extracted features can be visualized directly, or they can be used indirectly for modifying another type of visualization. In multifield data, each of the component fields can be searched for features, but in addition, there can be features of the multifield which rely on information form several of its components and which cannot be found by searching in a single field. In this chapter we give a survey of feature-based visualization of multifields, taking both of these feature types into account.

As with individual fields, one approach to visualizing multifields is to analyze the field and identify

features

. While some work has been carried out in detecting features in multifields, any discussion of multifield analysis must also identify techniques from single fields that can be extended appropriately.

Evaluation, solved and unsolved problems, and future directions are popular themes pervading the visualization community over the last decade. The top unsolved problem in both scientific and information visualization was the subject of an IEEE Visualization Conference panel in 2004. The future of graphics hardware was another important topic of discussion the same year. The subject of how to evaluate visualization returned a few years later. Chris Johnson published a list of 10 top problems in scientific visualization research. This was followed up by report of both past achievements and future challenges in visualization research as well as financial support recommendations to the National Science Foundation (NSF) and National Institute of Health (NIH). Chen recently published the first list of top unsolved information visualization problems. Future research directions of topology-based visualization was also a major theme of a workshop on topology-based methods. Laramee and Kosara published a list of top future challenges in human-centered visualization.

Similar to all other areas of visualization, visualization in biology and medicine is driven to a large extent by developments in the application domain itself. In recent years, new experimental techniques have increased measured data by orders of magnitude.

Connectomics is a branch of neuroscience that attempts to create a

connectome

, i.e., a complete map of the neuronal system and all connections between neuronal structures. This representation can be used to understand how functional brain states emerge from their underlying anatomical structures and how dysfunction and neuronal diseases arise. We review the current state-of-the-art of visualization and image processing techniques in the field of connectomics and describe a number of challenges. After a brief summary of the biological background and an overview of relevant imaging modalities, we review current techniques to extract connectivity information from image data at macro-, meso- and microscales. We also discuss data integration and neural network modeling, as well as the visualization, analysis and comparison of brain networks.

Basic biological research spans a huge range of scales, from studying the genome up to studying populations of people. Biological data is just as expansive. Advances in measuring devices and the public dissemination of large amounts of data has fundamentally changed the way that biologists conduct research and make scientific discoveries. Access to this data has made visualization a key component in almost every biological discovery workflow. In this chapter we focus on visualization in just a few areas of biology and highlight the challenges inherent within each. We present case studies from our own work to illustrate the impact that thoughtfully designed visualization systems can have on complex biological problems and highlight challenges for visualization research in these areas.

Due to continuing advances in medical imaging technology, and in medicine itself, techniques for visualizing medical image data have become increasingly important. In this chapter, we present a brief overview of the past 30 years of developments in medical visualization, after which we discuss the research challenges that we foresee for the coming decade.

Radiology is one of the main tools in modern medicine. A numerous set of deceases, ailments and treatments utilize accurate images of the patient. Ultrasound is one of the most frequently used imaging modality in medicine. The high spatial resolution, its interactive nature and non-invasiveness makes it the first choice in many examinations. Image interpretation is one of ultrasound’s main challenges. Much training is required to obtain a confident skill level in ultrasound-based diagnostics. State-of-the-art graphics techniques is needed to provide meaningful visualizations of ultrasound in real-time. In this paper we present the process-pipeline for ultrasound visualization, including an overview of the tasks performed in the specific steps. To provide an insight into the trends of ultrasound visualization research, we have selected a set of significant publications and divided them into a technique-based taxonomy covering the topics pre-processing, segmentation, registration, rendering and augmented reality. For the different technique types we discuss the difference between ultrasound-based techniques and techniques for other modalities.

Morphology of cardiovascular tissue is influenced by the unsteady behavior of the blood flow and vice versa. Therefore, the pathogenesis of several cardiovascular diseases is directly affected by the blood-flow dynamics. Understanding flow behavior is of vital importance to understand the cardiovascular system and potentially harbors a considerable value for both diagnosis and risk assessment. The analysis of hemodynamic characteristics involves qualitative and quantitative inspection of the blood-flow field. Visualization plays an important role in the qualitative exploration, as well as the definition of relevant quantitative measures and its validation. There are two main approaches to obtain information about the blood flow: simulation by computational fluid dynamics, and in-vivo measurements. Although research on blood flow simulation has been performed for decades, many open problems remain concerning accuracy and patient-specific solutions. Possibilities for real measurement of blood flow have recently increased considerably by new developments in magnetic resonance imaging which enable the acquisition of 3D quantitative measurements of blood-flow velocity fields. This chapter presents the visualization challenges for both simulation and real measurements of unsteady blood-flow fields.

In this chapter, we provide a brief overview of the visualization of large vector fields on parallel architectures using integration-based methods. After briefly providing background, we describe the state of the art in corresponding research, focusing on parallel integral curve computation strategies. We analyze the relative benefits of two fundamental schemes and discuss algorithmic improvements presented recently. To conclude, we point out open problems and future research directions.

As data sets grow in size and complexity, global analysis methods do not necessarily characterize the phenomena of interest, and scientists are increasingly reliant on feature-based analysis methods to study the results of large-scale simulations. This chapter presents a framework that efficiently encodes the set of all possible features in a hierarchy that is augmented with attributes, such as statistical moments of various scalar fields. The resulting meta-data generated by the framework is orders of magnitude smaller than the original simulation data, yet it is sufficient to support a fully flexible and interactive analysis of the features, allowing for arbitrary thresholds, providing per-feature statistics, and creating various global diagnostics such as Cumulative Density Functions (CDFs), histograms, or time-series. The analysis is combined with a rendering of the features in a linked-view browser that enables scientists to interactively explore, visualize, and analyze data resulting from petascale simulations. While there exist a number of potential feature hierarchies that can be used to segment the simulation domain, we provide a detailed description of two: the merge tree and the Morse-Smale (MS) complex, and demonstrate the utility of this new framework in practical settings.

As computational and experimental science have evolved, a new

dimension

of challenges for visualization and analysis has emerged: enabling research, understanding, discovery at multiple problem scales and the interaction of the scales, and abstractions of phenomena. Visualization and analysis tools are needed to enable interacting and reasoning at multiple simultaneous scales of representations of data, systems, and processes. Moreover, visualization is crucial to help scientists and engineers understand the critical processes at the scale boundaries through the use of external visual cognitive artifacts to enable more natural reasoning across these boundaries.

In computer science in general and in particular the field of high performance computing and supercomputing the term

scalable

plays an important role. It indicates that a piece of hardware, a concept, an algorithm, or an entire system scales with the size of the problem, i.e., it can not only be used in a very specific setting but it’s applicable for a wide range of problems. From small scenarios to possibly very large settings. In this spirit, there exist a number of fixed areas of research on

scalability

. There are works on scalable algorithms, scalable architectures but what are

scalable devices

? In the context of this chapter, we are interested in a whole range of display devices, ranging from small scale hardware such as tablet computers, pads, smart-phones etc. up to large tiled display walls. What interests us mostly is not so much the hardware setup but mostly the visualization algorithms behind these display systems that scale from your average smart phone up to the largest gigapixel display walls.

Although the amount and variety of data being generated is increased dramatically, the capabilities of data visualization, analysis, and discovery solutions have not been improved accordingly with the explosive rate of data production. One reason is that storage and processing at the level of raw data require supercomputer scale resources. The other is that working at the level of raw data prevents effective human comprehension while exploring and solving most problems. Here we show several approaches to scalable functional representations. Encoding, abstraction, and analysis at multiple scales of representations are a common approach in many scientific disciplines and provides a promising approach to harness our expanding digital universe.

With the ever-increasing capacity of high performance computing (HPC) systems, the computational simulation models become still finer and more accurate. However, the size and complexity of the data produced poses tremendous challenges for the visualization and analysis task. Especially when explorative approaches are demanded, distributed and parallel post-processing architectures have to be developed in order to allow interactive human-computer interfaces. Such infrastructures can also be exploited for the evaluation of ongoing simulation runs. The application here ranges from online monitoring to computational steering. But also remote and parallel rendering can be integrated into the overall setup. This chapter gives an overview of current solutions and ongoing research activities in this domain.