As visualization is applied to larger data sets residing in more diverse hardware environments, visualization frameworks need to adapt. Rendering techniques are currently a major limiter since they tend to be built around central processing with all of the geometric data present. This is not a fundamental requirement of information visualization. This paper presents Abstract Rendering (AR), a technique for eliminating the centralization requirement while preserving some forms of interactivity. AR is based on the observation that pixels are fundamentally bins, and that rendering is essentially a binning process on a lattice of bins. By providing a more flexible binning process, the majority of rendering can be done with the geometric information stored out-of-core. Only the bin representations need to reside in memory. This approach enables: (1) rendering on large datasets without requiring large amounts of working memory, (2) novel and useful control over image composition, (3) a direct means of distributing the rendering task across processes, and (4) high-performance interaction techniques on large datasets. This paper introduces AR in a theoretical context, provides an overview of an implementation, and discusses how it has been applied to large-scale data visualization problems.

With the current widespread interest in head mounted displays, we perceived a need for devices that support expressive and adaptive interaction in a low-cost, eyes-free manner. Leveraging rapid prototyping techniques for fabrication, we have designed and manufactured a variety of panels that can be overlaid on multi-touch tablets and smartphones. The panels are coupled with an app running on the multi-touch device that exchanges commands and state information over a wireless network with the virtual reality application. Sculpted features of the panels provide tactile disambiguation of control widgets and an onscreen heads-up display provides interaction state information. A variety of interaction mappings can be provided through software to support several classes of interaction techniques in virtual environments. We foresee additional uses for applications where eyes-free use and adaptable interaction interfaces can be beneficial.

The fundamental premise of visualization is that a useful correspondence between pixels and data can be built. However, visualization programs rarely operate at the pixel level. Instead of the discrete, finite space of pixels, the most common visualization models work with canvases of floating-point coordinates and geometric shapes. Preserving some source data link all the way down to the pixel level provides many opportunities for improving visualization. This paper describes Abstract Rendering (AR), a framework that preserves the data-to-pixel link. Using this pixel-level link, AR is used to provide a unique control over the final visual representation of data sets at all scales and from a variety of visualization technique families.

Rendering: Out-of-core Rendering for Information Visualization Joseph A. Cottama and Andrew Lumsdainea and Peter Wangb aCREST/Indiana University, Bloomington, IN, USA; bContinuum Analytics, Austin, TX, USA

It is impossible to directly visualize all of the items of a large dataset at once. Often, the number of items exceeds the number of pixels. Since direct representation is not a reliable option, a variety of methods have been developed for dealing with indirect representation. Such methods include clustering and intelligent filtering to reduce the number of items being considered in the first place. However, these techniques impose a high computational and interpretation costs. The alternative is to employ techniques to directly deal with the over-plotting that occurs. that occurs when there are too many items to display without overlapping. Over-plotting techniques include alpha composition, color weaving and selective plotting. Each of these has variants that yield different cognitive or computational optimizations. Unfortunately, most advanced over-plotting techniques are wrapped up in specific libraries. Experimenting with different techniques is cumbersome because they have not been provided with uniform interfaces or in a single runtime. This paper presents Abstract Rendering, a recasting of the rendering process that enables concise expression of many over-plotting techniques. Furthermore, the Abstract Rendering formulation yields efficient execution strategies. Combined, it is practical to explore different over-plotting techniques for large data without requiring significant alteration to existing pipelines.