FP-AmandaAlvarez

From CS294-10 Visualization Fa07

Jump to: navigation, search

Contents

[edit] Proposal

[edit] Group Members

  • Amanda Alvarez

[edit] Description

This project addresses the perception of motion in depth, in particular how percepts differ when viewing a real world physical movement and a stimulus moving on a display. Animation and motion in visualization can be compelling and informative factors, but there is no definitive account of which visual cues are important in the perception of motion. Investigating the cue space is particularly salient in the case of information visualization displays, since most flat displays provide a cue-inconsistent viewing environment. Pictorial depth cues on a flat screen induce a vergence-accommodation conflict in the viewer's eyes. When viewing a physical stimulus in the real world, this conflict does not exist, as objects at different vergence distances are actually also located in different planes.

In this study, we will investigate the perception of motion by comparing a real world viewing situation with a virtual scene on a display. In both cases, subjects will view an elliptical motion path head on, and will judge whether the path is too long or wide to be a circle. In the real world case, we have a mechanism similar to a planetary motion model that can display different elliptical paths. (Subjects only see the X and Z dimensions of motion; there is no movement in the Y, or up, dimension.) This task will be replicated on a 3D volumetric display, where the cue space can be manipulated. This display has three focal planes, and by using only one we can mimic the viewing situation with a conventional display. By using all three focal planes, we can reduce the cue inconsistency and make the vergence and accommodation cues agree with each other. In both the real and virtual cases, we can also manipulate other cues, such as looming (size cue), focus cues alone (using a pinhole), etc.

The relevance of this work to visualization is apparent when considering data visualizations that use 'rocking' to present more than two dimensions of data in a scatterplot. The motion that is supposed to inform the viewer about the data could be made much more informative if the motion were displayed in its full cue richness. Motion is an important cue to depth and structure, and it is exploited to enhance the percept of these aspects in visualizations.


References:

[edit] Initial Problem Presentation

[edit] Midpoint Design Discussion

[edit] Final Deliverables



[add comment]