| Summary: | The present study investigated depicted 3D orientation in texture gradients as a potential feature
to represent data attributes in scientific visualization applications. The task was to detect and/or
localize texture elements that varied across two components of 3D orientation, theta (left-right
orientation differences from y-axis) and phi (rotation about the y-axis). Experiment 1 determined
observers' detection sensitivity to orientation-defined 3D targets, and revealed that detection
performance for 3D targets was based on depicted 3D orientation rather than 2D image
orientation. In Experiment 2, the detection task was followed by a localization via direct action
(pointing to an onscreen location) or indirect action (making a spatially mapped keypress).
Results showed that if targets were correctly detected, they were also localized to the visual field
quadrant, with accuracy generally being higher for direct action responses. Additionally,
observers showed an ability to localize targets that they failed to detect. Experiment 3
demonstrated that, compared to performance of the component tasks alone, detection and direct
localization did not suffer in a dual-task context, but indirect localization did. These results
support a dual visual systems theory that postulates separate processing streams that specialize in
processing visual information for action or perception. Results also have practical implications
for how 3D orientation can be used in visualization applications.
|
|---|
