IEEE VIS Publication Dataset

We present the H3 layout technique for drawing large directed graphs as node-link diagrams in 3D hyperbolic space. We can lay out much larger structures than can be handled using traditional techniques for drawing general graphs because we assume a hierarchical nature of the data. We impose a hierarchy on the graph by using domain-specific knowledge to find an appropriate spanning tree. Links which are not part of the spanning tree do not influence the layout but can be selectively drawn by user request. The volume of hyperbolic 3-space increases exponentially, as opposed to the familiar geometric increase of euclidean 3-space. We exploit this exponential amount of room by computing the layout according to the hyperbolic metric. We optimize the cone tree layout algorithm for 3D hyperbolic space by placing children on a hemisphere around the cone mouth instead of on its perimeter. Hyperbolic navigation affords a Focus+Context view of the structure with minimal visual clutter. We have successfully laid out hierarchies of over 20,000 nodes. Our implementation accommodates navigation through graphs too large to be rendered interactively by allowing the user to explicitly prune or expand subtrees.

The Table Lens, focus+context visualization for large data tables, allows users to see 100 times as many data values as a spreadsheet in the same screen space in a manner that enables an extremely immediate form of exploratory data analysis. In the original Table Lens design, data are shown in the context area using graphical representations in a single pixel row. Scaling up the Table Lens technique beyond approximately 500 cases (rows) by 40 variables (columns) requires not showing every value individually and thus raises challenges for preserving the exploratory and navigational ease and power of the original design. We describe two design enhancements for introducing regions of less than a pixel row for each data value and discuss the issues raised by each.

Managing large projects is a very challenging task requiring the tracking and scheduling of many resources. Although new technologies have made it possible to automatically collect data on project resources, it is very difficult to access this data because of its size and lack of structure. We present three novel glyphs for simplifying this process and apply them to visualizing statistics from a multi-million line software project. These glyphs address four important needs in project management: viewing time dependent data; managing large data volumes; dealing with diverse data types; and correspondence of data to real-world concepts.

Metrics for information visualization will help designers create and evaluate 3D information visualizations. Based on experience from 60+ 3D information visualizations, the metrics we propose are: number of data points and data density; number of dimensions and cognitive overhead; occlusion percentage; and reference context and percentage of identifiable points.

The display of multivariate datasets in parallel coordinates, transforms the search for relations among the variables into a 2-D pattern recognition problem. This is the basis for the application to visual data mining. The knowledge discovery process together with some general guidelines are illustrated on a dataset from the production of a VLSI chip. The special strength of parallel coordinates is in modeling relations. As an example, a simplified economic model is constructed with data from various economic sectors of a real country. The visual model shows the interelationship and dependencies between the sectors, circumstances where there is competition for the same resource, and feasible economic policies. Interactively, the model can be used to do trade-off analyses, discover sensitivities, do approximate optimization, monitor (as in a process) and provide decision support.

We introduce nonlinear magnification fields as an abstract representation of nonlinear magnification, providing methods for converting transformation routines to magnification fields and vice-versa. This new representation provides ease of manipulation and power of expression. By removing the restrictions of explicit foci and allowing precise specification of magnification values, we can achieve magnification effects which were not previously possible. Of particular interest are techniques we introduce for expressing complex and subtle magnification effects through magnification brushing, and allowing intrinsic properties of the data being visualized to create data-driven magnifications.

This paper describes the SHriMP visualization technique for seamlessly exploring software structure and browsing source code, with a focus on effectively assisting hybrid program comprehension strategies. The technique integrates both pan+zoom and fisheye-view visualization approaches for exploring a nested graph view of software structure. The fisheye-view approach handles multiple focal points, which are necessary when examining several subsystems and their mutual interconnections. Source code is presented by embedding code fragments within the nodes of the nested graph. Finer connections among these fragments are represented by a network that is navigated using a hypertext link-following metaphor. SHriMP combines this hypertext metaphor with animated panning and zooming motions over the nested graph to provide continuous orientation and contextual cues for the user. The SHriMP tool is being evaluated in several user studies. Observations of users performing program understanding tasks with the tool are discussed.

Research on information visualization has reached the point where a number of successful point designs have been proposed and a variety of techniques have been discovered. It is now appropriate to describe and analyze portions of the design space so as to understand the differences among designs and to suggest new possibilities. This paper proposes an organization of the information visualization literature and illustrates it with a series of examples. The result is a framework for designing new visualizations and augmenting existing designs.

We describe a method for the visualization of information units on spherical domains which is employed in the banking industry for risk analysis, stock prediction and other tasks. The system is based on a quantification of the similarity of related objects that governs the parameters of a mass-spring system. Unlike existing approaches we initialize all information units onto the inner surface of two concentric spheres and attach them with springs to the outer sphere. Since the spring stiffnesses correspond to the computed similarity measures, the system converges into an energy minimum which reveals multidimensional relations and adjacencies in terms of spatial neighborhoods. Depending on the application scenario our approach supports different topological arrangements of related objects. In order to cope with large data sets we propose a blobby clustering mechanism that enables encapsulation of similar objects by implicit shapes. In addition, we implemented various interaction techniques allowing semantic analysis of the underlying data sets. Our prototype system IVORY is written in Java, and its versatility is illustrated by an example from financial service providers.

A method for efficiently volume rendering dense scatterplots of relational data is described. Plotting difficulties that arise from large numbers of data points, categorical variables, interaction with non-axis dimensions, and unknown values, are addressed by this method. The domain of the plot is voxelized using binning and then volume rendering. Since a table is used as the underlying data structure, no storage is wasted on regions with no data. The opacity of each voxel is a function of the number of data points in a corresponding bin. A voxel's color is derived by averaging the value of one of the variables for all the data points that fall in a bin. Other variables in the data may be mapped to external query sliders. A dragger object permits a user to select regions inside the volume.

The task of reconstructing the derivative of a discrete function is essential for its shading and rendering as well as being widely used in image processing and analysis. We survey the possible methods for normal estimation in volume rendering and divide them into two classes based on the delivered numerical accuracy. The three members of the first class determine the normal in two steps by employing both interpolation and derivative filters. Among these is a new method which has never been realized. The members of the first class are all equally accurate. The second class has only one member and employs a continuous derivative filter obtained through the analytic derivation of an interpolation filter. We use the new method to analytically compare the accuracy of the first class with that of the second. As a result of our analysis we show that even inexpensive schemes can in fact be more accurate than high order methods. We describe the theoretical computational cost of applying the schemes in a volume rendering application and provide guidelines for helping one choose a scheme for estimating derivatives. In particular we find that the new method can be very inexpensive and can compete with the normal estimations which pre-shade and pre-classify the volume (M. Levoy, 1988).

Triangle decimation techniques reduce the number of triangles in a mesh, typically to improve interactive rendering performance or reduce data storage and transmission requirements. Most of these algorithms are designed to preserve the original topology of the mesh. Unfortunately, this characteristic is a strong limiting factor in overall reduction capability, since objects with a large number of holes or other topological constraints cannot be effectively reduced. The author presents an algorithm that yields a guaranteed reduction level, modifying topology as necessary to achieve the desired result. In addition, the algorithm is based on a fast local decimation technique, and its operations can be encoded for progressive storage, transmission, and reconstruction. He describes the new progressive decimation algorithm, introduces mesh splitting operations and shows how they can be encoded as a progressive mesh. He also demonstrates the utility of the algorithm on models ranging in size from 1,132 to 1.68 million triangles and reduction ratios of up to 200:1.

Currently, the most popular method of visualizing music is music notation. Through music notation, an experienced musician can gain an impression of how a particular piece of music sounds simply by looking at the notes on paper. However, most listeners are unfamiliar or uncomfortable with the complex nature of music notation. The goal of this project is to present an alternate method for visualizing music that makes use of color and 3D space. This paper describes one method of visualizing music in 3D space. The implementation of this method shows that music visualization is an effective technique, although it is certainly not the only possible method for accomplishing the task. Throughout the course of this project, several variations and alternative approaches were discussed. The final version of this project reflects the decisions that were made in order to present the best possible representation of music data.

Previous accelerated volume rendering techniques have used auxiliary hierarchical datastructures to skip empty and homogeneous regions. Although some recent research has taken advantage of more efficient direct encoding techniques to skip empty regions, no work has been done to directly encode homogeneous but not empty regions. 3D distance transforms previously used to encode empty space can be extended to preprocess homogeneous regions as well, and these regions can be efficiently encoded and incorporated into volume ray-casting and back projection algorithms with a high degree of flexibility.

Splatting is a popular direct volume rendering algorithm. However, the algorithm does not correctly render cases where the volume sampling rate is higher than the image sampling rate (e.g. more than one voxel maps into a pixel). This situation arises with orthographic projections of high-resolution volumes, as well as with perspective projections of volumes of any resolution. The result is potentially severe spatial and temporal aliasing artifacts. Some volume ray-casting algorithms avoid these artifacts by employing reconstruction kernels which vary in width as the rays diverge. Unlike ray-casting algorithms, existing splatting algorithms do not have an equivalent mechanism for avoiding these artifacts. The authors propose such a mechanism, which delivers high-quality splatted images and has the potential for a very efficient hardware implementation.

An interactive cerebral blood vessel exploration system is described. It has been designed on the basis of neurosurgeons' requirements in order to assist them in the diagnosis of vascular pathologies. The system is based on the construction of a symbolic model of the vascular tree, with automatic identification and labelling of vessel bifurcations, aneurysms and stenoses. It provides several types of visualization: individual MRA (magnetic resonance angiography) slices, MIP (maximum intensity projection), shaded rendering, symbolic schemes and surface reconstruction.

In the area of scientific visualization, input data sets are often very large. In visualization of computational fluid dynamics (CFD) in particular, input data sets today can surpass 100 Gbytes, and are expected to scale with the ability of supercomputers to generate them. Some visualization tools already partition large data sets into segments, and load appropriate segments as they are needed. However, this does not remove the problem for two reasons: 1) there are data sets for which even the individual segments are too large for the largest graphics workstations, 2) many practitioners do not have access to workstations with the memory capacity required to load even a segment, especially since the state-of-the-art visualization tools tend to be developed by researchers with much more powerful machines. When the size of the data that must be accessed is larger than the size of memory, some form of virtual memory is simply required. This may be by segmentation, paging, or by paged segments. The authors demonstrate that complete reliance on operating system virtual memory for out-of-core visualization leads to egregious performance. They then describe a paged segment system that they have implemented, and explore the principles of memory management that can be employed by the application for out-of-core visualization. They show that application control over some of these can significantly improve performance. They show that sparse traversal can be exploited by loading only those data actually required.

This paper outlines a method to dynamically replace portals with textures in a cell-partitioned model. The rendering complexity is reduced to the geometry of the current cell thus increasing interactive performance. A portal is a generalization of windows and doors. It connects two adjacent cells (or rooms). Each portal of the current cell that is some distance away from the viewpoint is rendered as a texture. The portal texture (smoothly) returns to geometry when the viewpoint gets close to the portal. This way all portal sequences (not too close to the viewpoint) have a depth complexity of one. The size of each texture and distance at which the transition occurs is configurable for each portal.

Presents a new method for auralization of the vorticity of a streamline in a vector field. This technique involves using a composite tone formed by superimposing sine waves of various amplitudes whose frequency and amplitude vary in such a way as to give the perception that the resulting sound increases or decreases endlessly in pitch without ever extending beyond the listener's range of audible frequencies. Continuous clockwise or counterclockwise rotations of a streamline resulting from vorticity can then be displayed aurally as an apparently continuous increase or decrease in pitch.

Describes several visualization techniques based on the notion of multi-resolution brushing to browse large 3D volume datasets. Our software is implemented using public-domain libraries, and is designed to run on average-equipped desktop computers such as a Linux machine with 32 MBytes of memory. Empirically, our system allows scientists to obtain information from a large dataset with over 8.3 million numbers in interactive time. We show that very large scientific volume datasets can be accessed and utilized without expensive hardware and software.