Kenneth P. Herndon

Three-dimensional widgets

The 3D components of today’s user interfaces are still underdeveloped. Direct interaction with 3D objects has been limited thus far to gestural picking, manipulation with linear transformations, and simple camera motion. Further, there are no toolkits for building 3D user interfaces. We present a system which allows experimentation with 3D widgets, encapsulated 3D geometry and behavior. Our widgets are first-class objects in the same 3D environment used to develop the application. This integration of widgets and application objects provides a higher bandwidth between interface and application than exists in more traditional UI toolkit-based interfaces. We hope to allow user-interface designers to build highly interactive 3D environments more easily than is possible with today’s tools.

Aperture based selection for immersive virtual environments

We present two novel techniques for effectively selecting objects in immersive virtual environments using a single 6 DOF magnetic tracker. These techniques advance the state of the art in that they exploit the participant’s visual frame of reference and fully utilize the position and orientation data from the tracker to improve accuracy of the selection task. Preliminary results from pilot usability studies validate our designs. Finally, the two techniques combine to compensate for each other’s weaknesses.

Interactive shadows

It is often difficult in computer graphics applications to understand spatial relationships between objects in a 3D scene or effect changes to those objects without specialized visualization and manipulation techniques. We present a set of three-dimensional tools (widgets) called “shadows” that not only provide valuable perceptual cues about the spatial relationships between objects, but also provide a direct manipulation interface to constrained transformation techniques. These shadow widgets provide two advances over previous techniques. First, they provide high correlation between their own geometric feedback and their effects on the objects they control. Second, unlike some other 3D widgets, they do not obscure the objects they control.

The challenges of 3D interaction: a CHI '94 workshop

3D computer graphics is becoming more and more popular due to the increased availability of 3D hardware and software on all classes of computers. However, despite this growing popularity and the existence of a number of successful 3D graphics applications, particularly in CAD, CAE, and medical and scientific visualization, the field is still very immature, There are no widely accepted standards for hardware or software platforms; learning to implement or use 3D graphics software is still extremely laborious; and the most effective ways for humans to interact with synthetic 3D environments are still not clear.

The challenges of 3D interaction

Introduction Applications for viewing and manipulating 3D models and data are increasingly being used by the medical, scientific visualization, modeling and animation communities. The user interfaces to these applications have largely been 2D in nature, borrowing heavily from well-developed 2D user interface paradigms most commonly associated with windowing systems and GUI toolkits such as the Macintosh, Windows and Motif desktop toolkits. Recently, however, some research groups have begun to develop systems for designing and implementing new 3D user interface techniques specifically suited to 3D graphics applications. This emerging area of interface research stands to benefit greatly from organized discussion among researchers in a number of fields including 3D interactive graphics, perceptual psychology and cognitive science.

Using deformations to explore 3D widget design

We are developing a framework for creating interactive 3D environments for applications in design, education, and the communication of information and ideas [3]. Our most recent work focuses on providing a useful and powerful interface to such a complex environment. To this end we have developed 3D widgets, objects that encapsulate 3D geometry and behavior, to control other objects in the scene [2]. We build 3D widgets as first-class objects in our real-time animation system. Because our system allows rapid prototyping of objects, we hope to enlarge today’s surprisingly small vocabulary of 3D widgets that includes menus floating in 3D, gestural picking, translation and rotation, cone trees, and perspective walls. As a way to focus on issues of 3D widget design, we have developed widgets to perform a particular task: applying high-level deformations to 3D objects [1]. The complexity of these operations makes numerical specification or panels of sliders difficult to use, and yet direct manipulation interfaces cannot provide meaningful feedback without fixing most parameters. In this video paper, we show a set of new 3D widgets to control deformations called racks. A simple rack consists of a bar specifying the axis of deformation and some number of handles attached to the bar specifying additional deformation parameters. For example, a taper rack has two additional handles. Moving the ends of the handles towards or away from the axis bar changes the amount of taper of the deformed object; changing the distance between the handles changes the region over which the deformation is applied. A more complex rack can have multiple handles specifying different deformations. The racks in Figures 1–3 all have handles for twisting (purple), tapering (blue), and bending (red) an object. The deformation range is the region between the twist and taper handles. 2 The Issues in 3D Widget Design

An interactive 3D toolkit for constructing 3D widgets

Today’s user interfaces for most 3D graphics applications still depend heavily on 2D GUIs and keyboard input. There have been several recent attempts both to extend these user interfaces into 3D and to describe intermediary 3D widgets that control application objects [3; 4; 5; 7; 13; 15]. Even though this style of interaction is a straightforward extension of interaction through intermediary 2D widgets such as dials or sliders, we know of no efforts to develop interactive 3D toolkits akin to UIMX or Garnet [11]. The Brown Graphics Group has had considerable experience using its Unified Graphics Architecture (UGA) system [16] to script 3D widgets such as deformation racks [14], interactive shadows [9], parameterized models, and other constrained 3D geometries. Using this experience, we have developed an interactive toolkit to facilitate the visual programming of the geometry and behavior of such interactive models. The toolkit provides both a core set of 3D widget primitives for constructing interactive behaviors based on constrained affine transformations, and an interactive 3D interface for combining these primitives into more complex widgets. This video paper describes the fundamental concepts of the toolkit and its core set of primitives. In particular, we describe (i) the conceptual structure of the primitives, (ii) the criteria used to select a particular primitive widget set that would be expressive enough to let us construct a wide range of interactive 3D objects, and (iii) the constraint relationships among the primitives.

3D widgets for exploratory scientific visualization

Scientists use a variety of visualization techniques to help understand computational fluid dynamics (CFD) datasets, but the interfaces to these techniques are generally two-dimensional and therefore are separated from the 3D view. Both rapid interactive exploration of datasets and precise control over the parameters and placement of visualization techniques are required to understand complex phenomena contained in these datasets. In this paper, we present work in progress on a 3D user interface for exploratory visualization of these datasets.