Tutorial 1 - Sunday Morning

Introduction to Level Set Methods, Data Structures and Efficient Algorithms

Level: Advanced

Ken Museth, Linköping University & Digital Domain (Organizer)
Aaron Lefohn, Neoptica

The level set method is a mathematical tool for dynamic interface problems and as such has gained widespread popularity in an impressive array of applications ranging from segmentation in medical imaging to deforming surfaces in the general field of computer graphics. While numerous research papers and even a few text books have been devoted to this mathematically advanced method, very little material has been published on how to efficiently implement algorithms and data structures that form the backbone of this method. Additionally, people often find the learning curve for this method to be very steep since it involves unfamiliar mathematics. Consequently, many people either shy away from this powerful method or abandon it at a later stage when they realize that most applications are seriously limited by the underlying implementations. This is truly a shame considering the versatility and robustness of this technique, and it is the hope that the proposed course will bridge this gab in the literature and allow researchers and software developers to engage in new challenging level set applications. Specifically, the authors of this course have conducted research on level set applications and efficient implementations since year 2000 which has lead to numerous research papers, presentations and courses on this topic. We have had a tremendous amount of fun during this period and we hope to share the enthusiasm with as many as possible.

More specifically this course will focus on the mathematical and computational background material needed to fully understand and appreciated the potential of the level set method. As such it supplements the other course proposal focusing on level set application in visualization. We strongly feel that these two courses mutually supporting each other.

Tutorial 2 - Sunday Morning

Color in Information Display

Level: Beginner/Intermediate

Maureen Stone, StoneSoup Consulting (Organizer)

Color is a key component of information display that is easy to use badly. As a result, Edward Tufte's key principle for color design is "do no harm." The principles that underlie good color design have their roots in human perception, and a deep understanding of the color properties of different media. To be effective, color design, like all design, grows out of a clear model of the underlying information and its relationship to the user's task. This course will present the visual principles that inform good design, and the advances in color science and color technology that can be applied to the problem of using color effectively in information display.

Tutorial notes may be obtained by clicking here.

Tutorial 3 - Sunday Morning

Experimental Design and Analysis for Human-Subject Visualization Experiments

Level: All

J. Edward Swan II, Mississippi State University (Organizer)

This half-day tutorial is for researchers and engineers, working in the field of visualization, who wish to either (1) conduct visualization evaluation experiments with human subjects, and/or (2) gain a better understanding of the basic terminology of experimental design and analysis (e.g., the precise meaning of statements such as F(2,45) = 5.67, p =.023). To illustrate the discussed principles and techniques, the tutorial will include illustrative case studies of actual human-subject experiments conducted by the proposer. This tutorial will introduce the basics of experimental design and analysis. Especially regarding experimental analysis, it will focus on the fundamental logic behind topics such as hypothesis testing and analysis of variance, while avoiding the complexities that come from considering individual statistical tests.

Tutorial notes may be obtained by clicking here.

Tutorial 4 - Sunday Afternoon

Level Set Applications for Visualization

Level: Intermediate

David Breen, Drexel University (Organizer)
Ken Museth, Linköping University
Aaron Lefohn, Neoptica
Mikael Rousson, Siemens Corporate Research

Level set methods define dynamic surfaces implicitly as the level set (iso-surface) of a sampled, evolving n-D function. This course is targeted for researchers interested in learning about the application of level set methods/models to visualization. The course material will include lectures on a variety of level set applications. The course will be taught at an intermediate level. Therefore attendees should have a working knowledge of calculus, linear algebra, computer graphics and geometric modeling. Some familiarity with differential geometry, differential equations, numerical computing and image processing is strongly recommended, but not required. The course will describe in detail level set methods for 3D morphing, contour-based surface reconstruction, a volume dataset segmentation framework, advanced segmentation techniques that utilize statistical shape models, piecewise smooth intensity models and ordered spatial dependencies. The course will close with a lecture on interactive segmentation with level set models on GPUs.

Tutorial notes may be obtained by clicking here.

Tutorial 5 - Sunday Afternoon

Perceptual Issues for Visualization and Evaluation

Level: Beginner/Intermediate

Mark Livingston, Naval Research Laboratory (Organizer)

Fundamental aspects of human perception affect nearly all visualization systems. This tutorial discusses perceptual issues important for developing visualizations and evaluating the perceptual effectiveness of techniques. Visualization programmers may expect to become more aware of perceptual issues raised in their visual representations and be more able to identify confounding factors in designs for statistical studies of visualizations.

Discussions will use numerous case studies, demonstrations, and results from statistical studies to demonstrate surprising effects on human perception created by variables that are often used as parameters in visual representations. Topics will incorporate theories and conceptualizations of how various fundamental aspects of perception operate at a "user level." Areas will include spatial perception (notably depth), the appearance of color and numerous theories of how it operates, the perception of motion with a focus on how this affects animation, and on visual attention and operator workload in perceiving changes in visual displays.

Tutorial notes may be obtained by clicking here.

Tutorial 6 - Sunday Afternoon

GeoVisualization with Google Earth and GIS

Level: Beginner

Jason Dykes, City University, London (Organizer)
Jo Wood, City University, London

This short active tutorial has been designed to equip Information Visualization specialists with geo-spatial knowledge and skills that will help them use geographic information techniques and technologies effectively in their visualization.

Our 'hands on' introduction focuses on visualizing the population - a much-mapped phenomenon and a well known data type - through a series of accessible technologies and formats. We highlight a number of key issues in Geographic Information Science that are fundamental to effective geovisualization and aim to draw attention to key considerations in geovisualization, share expertise across disciplinary boundaries and develop cross-disciplinary dialogue.

Having completed the tutorial participants will be able to:

• specify spatial entities in KML for visualization in Google Earth
• visually encode cartographic symbolism in KML
• use KML to encode structure that results in interactions in Google Earth
• visualize continuous surfaces in a GIS
• use visualization to compare and contrast representations of geographic data
• use visualization to consider the effects of scale in geographic data

The tutorial will be organised as an 'active learning' session. Each of the major topics will be introduced by an instructor. This will be followed by time for participants to complete a short task that allows them to incrementally build a simple visualization application from diverse data sources.

Those participating will be expected to do so in a 'hands on' manner and a notebook computer will be required. Participants will be asked to install freely available software in advance of the tutorial. This will be made available for download from the Web in advance of the tutorial and subsequently.

No prior experience with using geographic information or GIS is assumed.

No programming skills are required.

Participants will be editing example KML and text files and should ensure that they have access to a text editor on their notebook.

Datasets, software downloads, and recent information can be found by clicking here.

Tutorial notes may be obtained by clicking here.

Tutorial 7 - Monday Morning

Illustrative Display and Interaction in Visualization

Level: All

Ivan Viola, University of Bergen (Organizer)
Stefan Bruckner, Vienna University of Technology
Mario Costa Sousa, University of Calgary
David S. Ebert, Purdue University
Carlos D. Correa, Rutgers University

The area of illustrative visualization is concerned with developing methods to enhance the depiction of scientific data based on principles founded in traditional illustration. The illustration community has century-long experience in adapting their techniques to human perceptual needs in order to generate an effective depiction which conveys the desired message. Thus, their methods can provide us with important insights into visualization problems.

In this tutorial, the concepts in illustrative visualization are reviewed. An important aspect here is interaction: while traditional illustrations are commonly only presented as static images, computer-assisted visualization enables interactive exploration and manipulation of complex scientific data. Only by coupling illustrative visualization with effective interaction techniques its full potential can be exploited.

The tutorial starts with a general introduction into the area of illustrative visualization. The concept of importance-driven visualization and its applications are presented. Then we proceed with discussion how traditional abstraction techniques can be applied in an interactive context using importance-based methods. This ranges from low-level appearance to smart viewpoint-dependent visibility techniques such as cutaways or exploded views.

Further advanced manipulation strategies are discussed in the third part. The use deformations to enhance visibility of certain features while providing context or to abstract the structure of a complex objects through direct interaction with the data is examined. As many of the presented methods rely on a separation of focus and context, i.e., the important structures in the data have been identified, the tutorial discusses approaches for selecting objects of interest in a three-dimensional environment using intuitive sketch-based interfaces. Since the effectiveness of a user-interface is heavily dependent on the previous knowledge of the user, the last part of the tutorial examines the concept of layering interfaces based on user expertise. Finally, the application of illustrative display and interaction techniques for non-traditional modalities such as mobile devices concludes the tutorial.

Tutorial notes may be obtained by clicking here.

Tutorial 8 - Monday Morning

Introduction to Visual Medicine: Techniques, Applications and Software

Level: Beginner

Bernhard Preim, University of Magdeburg (Organizer)
Dirk Bartz, University of Leipzig
Felix Ritter, MeVis - Center for Medical Diagnostic Systems and Visualization,
Klaus Müller, Stony Brook University
Karel Zuiderveld, Vital Images

One of the largest application domains of visualization is medicine. 3D visualization techniques are essential in daily health-care practice and in clinical research. In particular, recent developments in image acquisition, diagnostic procedures and minimally-invasive surgery require an advanced planning and intra-operative support through computer science methods. With the increased spatial and temporal resolution, 3D visualizations become important for selected diagnostic procedures and for a wide range of therapy planning scenarios ranging from radiation treatment planning to neurosurgical interventions. In this tutorial, we will give an introduction into medical imaging methods---such as data acquisition, data analysis, segmentation, registration and rendering. Surface and volume rendering (including transfer function specification) and hybrid combinations of both are covered. Acceleration strategies which provide efficient rendering without (significant) loss of accuracy and image quality are described.We will not only discuss the advantages to traditional methods, but also illustrate their specific and inherent limitations. Based on this foundation, we will describe how these techniques are used giving real-world examples.

Tutorial notes may be obtained by clicking here.

Tutorial 9 - Monday Afternoon

Scope and Challenges of Visual Analytics

Level: Beginner/Intermediate

Daniel Keim, Konstanz University (Organizer)
Jim Thomas, Pacific Northwest National Laboratory

The science of Visual Analytics is a relatively new multidisciplinary field that combines various research areas including visualization, human-computer interaction, data analysis, data management, geo- spatial and temporal data processing and statistics. The goal of Visual Analytics is derive insight from massive, dynamic, ambiguous, and often conflicting data; detect the expected and discover the unexpected; provide timely, defensible, and understandable assessments; and communicate the assessment effectively for action. Providing visual access to and enabling insightful understanding of large amounts of complex data will be a major driving force in many research and technological development areas. The better integration of the increasing processing power of computers with the knowledgeable perception of the human user is a challenging and promising road to solutions for the information overload. Goal of the tutorial is to provide an overview of Visual Analytics, defining the scope of research in this novel area, outlining the challenges of different applications, and presenting Visual Analytics techniques and systems. The tutorial will provide the necessary background to understand the issues in the development and usage of Visual Analytics techniques. We will provide a number of case studies and show samples of commercial and academic Visual Analytics systems.

Tutorial notes may be obtained by clicking here.

Tutorial 10 - Monday Afternoon

Advanced Visual Medicine: Techniques, Applications and Software

Level: Intermediate/Advanced

Steffen Oeltze, University of Magdeburg (Organizer)
Dirk Bartz, University of Leipzig
Florian Link, MeVis - Center for Medical Diagnostic Systems and Visualization
Bernhard Preim, University of Magdeburg
Anna Vilanova, Eindhoven University of Technology
Stefan Zachow, Zuse-Institute Berlin (ZIB)

One of the largest application domains of visualization is medicine. 3D visualization techniques are essential in both daily health-care practice and in clinical research. With the increased spatial and temporal resolution, 3D visualizations combined with simulation and visual analysis techniques become important for selected diagnostic procedures and for a wide range of therapy planning scenarios.

In this tutorial, we explore a variety of advanced topics of visual medicine. In particular, we will discuss the visualization of vascular structures, the exploration and visual analysis of perfusion data, diffusion tensor imaging (with applications in ischemic heart studies and neonate studies), fast tagged multi-resolution volume rendering, OR-fit mixed reality methods for surgery, and the integration of simulation and visualization for surgical planning some of the most actively researched fields in visual medicine. Besides the technical aspects, we will introduce the specific application problems and discuss the requirements on dedicated solutions. The tutorial includes software issues based on an advanced, freely available software system for visualization in medicine: MeVisLab.

Tutorial notes may be obtained by clicking here.

Tutorial 11 - Wednesday Morning

Bridging the Chasm between InfoVis Research and the World Out There

Level: All

Stephen Few, Perceptual Edge (Organizer)

A wide chasm continues to separate infovis researchers from the software vendors that could benefit from their work. Software companies (including most business intelligence vendors) are responsible for half of the chasm. With few exceptions, they know little about information visualization and are doing almost nothing to change this. Researchers, however, are responsible for the other half. This is largely due to three failures:

• Much infovis research has no practical application.
• Much infovis research produces incomprehensible visualizations and ineffective functionality.
• Much infovis research is not presented in an understandable and compelling manner.

Researchers who want their work to matter must respond to the actual needs of people. Pursuing what interests you is great if it happens to coincide with what's needed; otherwise, it might end up mattering to no one but you. Producing work that looks cool but is incomprehensible is a similar waste. Perhaps the greatest travesty, however, is great research that matters and works, but remains on the shelf because you've failed to present it in way that informs and inspires.

This tutorial is for infovis researchers who want their work to find its way into software where it can make a difference beyond the walls of academia.

Topics include:

• Overview of the current use of infovis in business applications
• How to determine what infovis research is most needed
• Fundamental principles of visual perception and how they apply to infovis
• What all infovis products should do to augment cognition
• Appropriate color choices
• Appropriate data encoding object choices
• Appropriate balancing of visual salience
• Elimination of distracting visual content and/or attributes
• Best practices in data presentation

Tutorial notes may be obtained by clicking here.

Tutorial 12 - Wednesday Afternoon

Statistical Graphics for High-D Data

Level: Intermediate

Debby Swayne, Iowa State University (Organizer)
Heike Hofmann, Iowa State University
Michael Lawrence, Iowa State University
Hadley Wickham, Iowa State University

In this tutorial we'll describe interactive statistical graphics methods for visualizing high-dimensional data. Graphics are a fundamental part of data analysis, used in initial data inspection and exploration, model building and checking and also communicating information. You will get the most out of the tutorial if you have struggled with visualizing high-dimensional data in the past. We will be presenting techniques from a different ancestry than most infovis tools, so expect to learn some new ideas.

We will teach the basics of direct manipulation and dynamic graphics that facilitate exploratory data analysis. We will describe multiple linked windows and grand, guided and manual tours. Direct manipulation on the plots includes scaling, moving points, linked brushing and identification using categorical variables. Visual methods for handling missing values, supervised classification, clustering and statistical inference for plots.

The methods taught are readily available in open source software, enabling all participants to reproduce, extend and use them with their own data after the workshop.