HCI Bibliography Home | HCI Conferences | GI Archive | Detailed Records | RefWorks | EndNote | Hide Abstracts
GI Tables of Contents: 868889909192939495969798990001020304050607

Proceedings of the 1997 Conference on Graphics Interface

Fullname:Proceedings of the 1997 Conference on Graphics Interface
Editors:Wayne A. Davis; Marilyn Mantei; R. Victor Klassen
Location:Kelowna, B.C., Canada
Dates:1997-May-21 to 1997-May-23
Publisher:Canadian Information Processing Society
Standard No:ISBN 0-9695338-6-1; hcibib: GI97
Papers:22
Pages:199
Links:Conference Series Home Page | Online Proceedings
  1. Animation
  2. Illumination
  3. Rendering
  4. Visualization
  5. Interaction and Design
  6. Representation
  7. Modeling

Animation

Simulation levels of detail for real-time animation BIBAGZHTML 1-8
  Deborah A. Carlson; Jessica K. Hodgins
When sufficient computing power is available, dynamic simulation can be used as a source of motion for real-time, interactive virtual environments. In this paper, we explore techniques for reducing the computational cost of simulating groups of creatures by using less accurate simulations for individuals when they are less important to the viewer or to the action in the virtual world. The less accurate, or lower level of detail, simulations can be dynamic with fewer degrees of freedom, hybrid kinematic/dynamic, or purely kinematic. As a test of the effectiveness of this approach, we implemented an environment with dynamically simulated legged creatures. Because the creatures switch smoothly among different levels of detail for the underlying simulation, we can achieve real-time performance for a larger group of creatures than would be possible if each creature were dynamically simulated. To be useful in this test case, the method must meet two criteria: the outcome of the game must be essentially unchanged and the viewer's perception of the motion must be the same. While it is not possible to make definitive measurements of these criteria, we assess the performance gain from using different levels of detail and make a preliminary assessment of the effect that the decreased accuracy has on the outcome of the sample game.
Soundtracks for computer animation: sound rendering in dynamic environments with occlusions BIBAGZ 9-16
  Nicolas Tsingos; Jean-Dominque Gascuel
With the development of virtual reality systems and multi-modal simulations, soundtrack generation is becoming an important issue in computer graphics. In the context of computer generated animation, many more parameters than the sole object geometry as well as specific events can be used to generate, control and render a soundtrack that fits the object motions.
   Producing a convincing soundtrack involves the rendering of the interactions of sound with the dynamic environment: in particular sound reflections and sound absorption due to partial occlusions, usually implying an unacceptable computational cost.
   We present an integrated approach to sound and image rendering in a computer animation context, which allows the animator to recreate the process of sound recording, while ''physical effects are automatically computed. Moreover, our sound rendering process efficiently combines a sound reflection model and an attenuation model due to scattering/diffraction by partial occluders, through the use of graphics hardware allowing interactive computation rates."
Do geometric models affect judgments of human motion? BIBAKGZHTML 17-25
  Jessica K. Hodgins; James F. O'Brien; Jack Tumblin
Human figures have been animated using a wide variety of geometric models including stick figures, polygonal models, and NURBS-based models with muscles, flexible skin, or clothing. This paper reports on experiments designed to ascertain whether a viewer's perception of motion characteristics is affected by the geometric model used for rendering. Subjects were shown a series of paired motion sequences and asked if the two motions in each pair were ''the same'' or ''different.'' The two motion sequences in each pair used the same geometric model. For each trial, the pairs of motion sequences were grouped into two sets where one set was rendered with a stick figure model and the other set was rendered with a polygonal model. Sensitivity measures for each trial indicate that for these sequences subjects were better able to discriminate motion variations with the polygonal model than with the stick figure model.
Keywords: motion perception, motion sensitivity, perceptual study, computer animation, polygonal model, geometric model

Illumination

An Image-Space Refinement Criterion for Linear Hierarchical Radiosity BIBAGZGZGZ 26-36
  Ignacio Martin; Xavier Pueyo; Dani Tost
A Linear Hierarchical Radiosity method using point collocation and triangle meshes is proposed that allows C0 continuity and performs energy exchange at any level both in the shooter and the receiver; this method leads to an exact representation of the Gouraud shading interpolation that will be used for rendering. A new refinement criterion is presented which tries to improve image quality taking into account: the smoothness of the solution based on pixel intensity values instead of energy ones, and visibility changes along the surfaces for high gradient detection (sharp shadows). In order to perform a efficient refinement a data structure is proposed which isolates every shooter contribution over each receiver and allows to only refine high error interactions.
Probability trees BIBAGZ 37-46
  Michael D. McCool; Peter K. Harwood
A k-D tree representation of probability distributions is generalized to support generation of samples from conditional distributions. An interpretation of the approach as a piecewise linear warping function is provided which permits a priori stratified sample generation. The representation is related to higher-order spline estimators and representations via projection. An application in glyph-based volume visualization is presented.
Properties of Tabulated Bidirectional Reflectance Distribution Functions BIBAGZ 47-55
  Joel DeYoung; Alain Fournier
One way to overcome the limitations imposed by analytical models of reflection is to use discretely sampled reflectance data directly. Through either empirical measurement or simulation, a bidirectional reflectance distribution function (BRDF) is sampled and stored as a table of numbers. The generality of these tabulated BRDFs is useful for generating realistic images, but the inevitable inaccuracy associated with the data gathering process can lead to a BRDF that is much more general than it needs to be, or that lacks certain physical properties.
   In this paper we propose measures for several properties of tabulated BRDFs: reciprocity, energy conservation, isotropy, and separability. Techniques to transform tabulated BRDFs to match one or more of these properties are also described. These transformations allow compression of the BRDF data, elimination of noise, improved computation time in some rendering tasks, and improved compliance with physical laws.
Light scattering simulations using complex subsurface models BIBACitation 56-67
  Morgan Schramm; Jay Gondek; Gary Meyer
New techniques are presented for simulating complex surface reflection phenomena. The approach involves the modeling of surface and subsurface microstructures, the simulation of how light interacts with this geometry, and the recording of the scattered light as a bidirectional reflection distribution function. Previous light simulation techniques employed in this method are extended to include subsurface refraction and multilayer interference effects. A complicated modeling and simulation example is presented that involves the structure of a hummingbird feather and multilayer interference. Light reflection measurements are recorded from paper samples, and a second modeling and simulation exercise is undertaken in an attempt to reproduce the measured results. In both cases, first order agreement is achieved between the simulations, the available analytical models, and the measured data.

Rendering

An image-based model for realistic lens systems in interactive computer graphics BIBAGZ 68-75
  Wolfgang Heidrich; Philipp Slusallek; Hans-Peter Seidel
Many applications, such as realistic rendering, virtual and augmented reality, and virtual studios, require an accurate simulation of real lens and camera systems at interactive rates, including depth of field and geometric aberrations, in particular distortions. Unfortunately, camera models used in Computer Graphics are either too simple to describe these effects or too expensive to simulate for interactive use.
   In this paper, we introduce an image-based lens model that is powerful enough to simulate sophisticated properties of real lens systems, yet fast enough for interactive graphics. By exploiting coherency, common graphics hardware can be used to yield high frame rates.
Directed safe zones and the dual extent algorithms for efficient grid traversal during ray tracing BIBAGZGZGZ 76-87
  Sudhanshu Kumar Semwal; Hakan Kvarnstrom
Ray tracing is inherently a very time consuming process. There have been a variety of techniques developed for reducing the rendering time. While using space subdivision techniques, the object space is divided into a set of disjoint voxels; and only those objects are checked for intersection with the ray which pierce the voxels along the path of the ray. As the grid size is increased in an attempt to reduce the number of objects encountered along the path of the ray, more empty voxels are encountered along the path. In other words, considerable rendering time could be invested in moving from one empty voxel to another empty voxel. The proximity clouds method uses distance transformations to identify the empty regions, and combined with the 3DDA grid traversal (SEADS) technique, has been shown to be the fastest grid traversal technique available today.
   In this paper, we present two further improvements to the proximity clouds method -- called the Directed Safe Zones (DSZ) and the Dual Extents (DEs). We have implemented four methods for our comparison: SEADS (3DDA), proximity clouds, Directed Safe Zones (DSZs), and Dual Extents (DEs). We present both the theoretical and statistical analysis of the four methods. We show that DSZs and DEs surpass the performance of both the proximity clouds and SEADS implementations. The DSZs method \em usually outperforms the Dual Extents implementation, except when the topology of the scene favors the Dual Extent method.
Multiresolution BSP trees applied to terrain, transparency, and general objects BIBAGZ 88-96
  Charles Wiley; A. T., III Campbell; Stephen Szygenda; Donald Fussell; Fred Hudson
We present a system for incorporating multiple level of detail (LOD) models of 3D objects within a single hierarchical data structure. This system was designed for a scientific visualization application involving terrain and volume rendering. Our data structure is a modified Binary Space Partitioning (BSP) tree. We describe how our tree construction and traversal routines may be used with a variety of LOD methods. This is demonstrated with two different LOD methods: a new method specialized for terrain elevation height fields, and an existing method for general objects. Images, timings, and storage data for our implementation are provided.
Multiresolution rendering of complex botanical scenes BIBAGZ 97-104
  Dana Marshall; Donald S. Fussell; A. T., III Campbell
This paper presents a system for rendering very large collections of randomly parameterized plants while generating manageable scene geometries for rendering. A given botanical description of a plant is compiled into a hierarchical volume approximation. This is then integrated into a multiresolution rendering system that uses adaptive volume refinement. For objects that are close to the viewer, explicit polygons are generated, while objects hidden or further away are rendered as groups of microsurfaces. This system can be extended to any polygon intensive rendering.

Visualization

Coherent zooming of illustrations with 3D-graphics and text BIBAGZ 105-113
  Bernhard Preim; Andreas Raab; Thomas Strothotte
In this paper we develop a method for zooming in complex information spaces consisting of geometric models and associated textual explanations. In particular we introduce a technique to extend fisheye views to the exploration of three-dimensional models.
   The result of a zooming process is a modified geometric model which is rendered and presented to the user with appropriately selected and well-placed textual labels. The user interacts with such coherent presentations to explore the information space.
   Our methods have been implemented in a system called the ZOOM ILLUSTRATOR. The system maintains a close relationship between images and associated text, with interaction on the textual part influencing the appearance of the graphical part and vice versa. We demonstrate our system on medical illustrations, in particular from anatomy.
Visual Metaphors for Understanding Logic Program Execution BIBAGZ 114-120
  Eric Neufeld; Anthony Kusalik; Michael Dobrohoczki
A classic notion in logic programming is the separation of logic and control. Logic is for problem solving; control is for directing inference. However, practical experience in the classroom suggests that problem-solving students nonetheless devote much effort to understanding control issues such as eliminating looping behaviours and improving program efficiency.
   In the case of Prolog, this requires a clear understanding of the operation of both unification and backtracking. Students often try to get this understand by tracing executions, but the common four-port debugger used in Prolog is not as helpful as it could be. In particular, it provides information in low bandwidth textual form.
   This paper describes a new visualization system for logic programming that uses colour tagging to trace unification through the Prolog proof tree. A user can dynamically tag" a term or subterm with a colour that is immediately propagrated through the displayed tree. The colour is also propagated through the proof tree on subsequent execution steps. This "colour unification" has an interesting relationship with usual Prolog unification. Initial experiences reveal several interesting visual metaphors that assist understanding of logic program execution. Experience to date also suggests new directions for visualization of logic programs."
Interactive Volumetric Information Visualization for Document Corpus Management BIBAGZ 121-128
  David S. Ebert; Chris D. Shaw; Amen Zwa; Ethan L. Miller; D. Aaron Roberts
This paper describes a minimally immersive three-dimensional volumetric interactive information visualization system for management and analysis of document corpora. The system, SFA, uses glyph-based volume rendering, enabling more complex data relationships and information attributes to be visualized than traditional 2D and surface-based visualization systems. Two-handed interaction using three-space magnetic trackers and stereoscopic viewing are combined to produce a minimally immersive interactive system that enhances the user's three-dimensional perception of the information space. This new system capitalizes on the human visual system's pre-attentive learning capabilities to quickly analyze the displayed information. We describe the usefulness of this system for the analysis and visualization of document similarity within a corpus of textual documents.

Interaction and Design

The immediate usability of graffiti BIBAGZ 129-137
  I. Scott MacKenzie; Shawn Zhang
We present four empirical measures of the immediate usability of Graffiti, a character recognizer for pen-based computers. Since speed is fully controlled by the user, we measured the accuracy attainable after minimal exposure. The first measure, 79%, is the inherent accuracy, or the extent to which Graffiti strokes match letters in the Roman alphabet. The other three measures were obtained in a formal experiment. We asked 25 subjects to enter the alphabet five times into a pen-based computer under three conditions: (a) following one minute studying the Graffiti reference chart, (b) following five minutes of practicing with Graffiti, and (c) following a one week lapse with no intervening practice. The accuracy was 86%, 97%, and 97%, respectively. These are very respectable figures given the limited exposure of subjects. The third figure represents complete retention following a one-week lapse. We present analyses of the errors on a character-by-character basis, noting that poor performing characters should be emphasized in tutorials and other learning aids for new users.
The Effect of Turn-Taking Protocols on Children's Learning in Mouse-Driven Collaborative Environments BIBAGZ 138-145
  Kori Inkpen; Joanna McGrenere; Kellogg S. Booth; Maria Klawe
This study compared the influence of turn-taking protocols on children's behaviour and learning when they used either one shared mouse or two individual mice in a collaborative problem-solving environment. The two-mouse case was investigated for both a give protocol, in which the child with control of the game voluntarily relinquished control, and a take protocol, in which the child without control of the game preemptively acquired control. Children in the study took part in two sessions. In the first collaborative session, children played a problem solving puzzle game with a partner using one of the three protocols (one- mouse shared, two-mouse give, or two-mouse take). This was followed by a second solo session in which each child played the game alone. The results of the study revealed that the choice of turn-taking protocol can have a significant affect on children's learning and behaviour in a collaborative problem-solving environment. For boys, the protocol affected their access to the mouse, which in turn affected their learning: a significant correlation was found between the amount of time each boy had control of the mouse in the collaborative session and the number of puzzles that same boy could solve in a subsequent solo session. In the two-mouse take condition boys exhibited a more equal division of mouse control than did boys using either of the other two protocols.
World space surface pasting BIBAGZ 146-154
  Leith Kin Yip Chan; Stephen Mann; Richard Bartels
Surface pasting is a composition method that applies features to base surfaces to provide details on the base surfaces. The location and size of a feature are determined by the transformations of the feature's domain. By modifying the domain layout of pasted surfaces, we can manipulate the appearance of a feature interactively in a Domain Space User Interface. However, this domain user interface is inconvenient because the user cannot interact with the three-dimensional model directly. In this paper, we describe a World Space User Interface that maps actions on the three space surfaces to two-dimensional domain operations.

Representation

Linear fractal shape interpolation BIBAGZ 155-162
  Brandon Burch; John C. Hart
Interpolation of two-dimensional shapes described by iterated function systems is explored. Iterated function systems define shapes using self-transformations, and interpolation of these shapes requires interpolation of these transformations. Polar decomposition is used to avoid singular intermediate transformations and to better simulate articulated motion. Unlike some other representations, such as polygons, shaped described by iterated function systems can become totally disconnected. A new, fast and image-based technique for determining the connectedness of an iterated function system attractor is introduced. For each shape interpolation, a two parameter family of iterated function systems is defined, and a connectedness locus for these shapes is plotted, to maintain connectedness during the interpolation.
A heuristic method for generating 2D CSG trees from bitmaps BIBAKGZ 163-172
  Sarah F. Gibson; Joe Marks; Danielle Feinberg; Manuel Sosa
This paper presents a method for generating compact and effective constructive-solid-geometry (CSG) object representations from 2D bitmap representations. The method consists of two parts: a stochastic search procedure for finding candidate CSG trees and a local optimization procedure that modifies the primitives of a given CSG tree to effect a better match. Results for several sample input bitmaps are presented and an application of this method to automatic geometric morphing between pairs of bitmaps is shown.
Keywords: constructive solid geometry, conversion, optimization, geometric morphing

Modeling

An Efficient Volumetric Method for Building Closed Triangular Meshes from 3-D Image and Point Data BIBAGZGZ 173-180
  Gerhard Roth; Eko Wibowoo
We present a volumetric method that can efficiently create triangular meshes from 3-D geometric data. This data can be presented in the form of images, profiles or unordered points. The mesh model can be created at different resolutions and can also be closed to make a true volumetric model.
An interactive fur modeling technique BIBAGZ 181-188
  Allen Van Gelder; Jane Wilhelms
A technique for modeling fur, but not long human hair, quickly, using the facilities of common graphics workstations is described. The user selects a variety of parameters to achieve the desired appearance for a particular animal, such as hair density, length, stiffness, and color properties. Undercoat and overcoat may have separate specifications, and degrees of randomness may be specified, for added realism. Standard GL facilities are used for modeling, lighting, and rendering. Hair density is automatically adjusted for viewing distance to compensate for the limit of single pixel resolution, thus avoiding the tendency to ''melt'' into a surface of uniform appearance. Gravitational effects are approximated. Four drawing methods are compared: single line, polyline, nurbs curve, and nurbs surface. The polyline method is judged to offer reasonable realism at substantially faster rendering rates than previously reported hair techniques.
A modelling method and user interface for creating plants BIBAGZ 189-197
  Oliver Deussen; Bernd Lintermann
We present a modelling method and graphical user interface for the creation of natural branching structures such as plants. Structural and geometric information is encapsulated in objects that are combined to form a description of the model. The description is represented graphically as an icon tree and can be edited interactively.
   Global and partial constraint techniques are integrated on the basis of tropisms and allow the modelling of specific shapes. We show examples to illustrate the design process and evaluate the user interface.