Spatial Cues in 3D Visualization - Ambient Intelligence for Scientific Discovery - page 104

Spatial Cues in 3D Visualization

Geoffrey S. Hubona 1 and Gregory W. Shirah 2

1 Georgia State University, USA
[e-mail]
2 NASA Goddard Space Flight Center, USA
[e-mail]

1 Introduction

The ever-increasing power and complexity of available hardware and software
has enabled the development of a wide variety of visualization techniques that
allow the ever more concise presentation of data. Associated with this trend is the
challenge to condense and convey ever-increasing amounts of useful information
into smaller and smaller spaces. Depicting computer-generated visualizations
in three dimensions (3D), similar to how we perceive the real world, is one
approach to condense these visual presentations of information. However, there
is an inherent dilemma in this approach; the visual medium on which the vast
majority of 3D imagery is displayed is inherently two dimensional (2D): a flat
computer monitor. Although there are some immersive or 'true 3D devices' on
the market, such as the fishbow rotating display, the LCD layered 3D monitor,
and 3D displays marketed by companies such as SeeReal Technologies, most of
these devices are either too small or too expensive for the average user.
Humans visually discover and interact with their environment by: (1) perceiv-
ing visual information; (2) recognizing the “external source” of the information;
and (3) interpreting its meaning and significance. The process of human percep-
tion and the interpretation of visual information is extremely complex, involving
many levels of anatomical, neurochemical and psychophysical processing in vi-
sion and cognition. The human eye does not operate like a camera, but rather
serves as an 'optical interface' between the external environment and the internal
neural components of human vision and cognition which convey the basic visual
attributes of form, field, color, motion, and depth. It is the design of computer
visualizations to promote the accurate understanding of the relative depths (e.g.
in the 'z' dimension) of objects that is particularly challenging.
The visual cues that enable us to perceive depth have been extensively doc-
umented. However, the use of these cues, and the combination of these cues, to
effectively convey depth information in computer-generated imagery is still an
ongoing topic of contemporary research. Depth cues are often placed into two
categories: (1) primary depth cues, with a physiological basis, that include binoc-
ular disparity (which enables stereopsis), convergence and accommodation; and
(2) secondary (or pictorial) cues, including texture and texture gradients, shad-
ing and shadow, relative motion, occlusion, reference frames, linear and aerial
perspective, height in plane (or elevation), apparent size, and others (please see
[1] and [2] for a complete discussion of primary and secondary depth cues).