Shape in glyph-based visualization

Visualization of continuous data using shape in glyph-based visualization.

Pradyut Panda,
Department of Computer Science,
University of Maryland, Baltimore County.
Advisor: Dr. David Ebert.
15th July 1998.


This paper looks at the use of shape to represent or visualize a single dimension of continuously varying data in a glyph-based rendering system that is used for the visualization of multi-variate datasets. We use procedurally generated shapes to generate the shape of the glyph. Three algo rithms were used on this paper to generate the shape of the glyph. The first method was the distor tion of a basic shape such as a cube using a fractional Brownian motion type of turbulence function. At low values the distortion is minimal and the shape closely resemble it's basic shape whereas at high values the shape is highly distorted. However it is possible to see the underlying basic shape. The second technique to generate shape that we looked at, was the gradual warping from one pre-defined shape to another. This was achieved by performing a linear interpolation between corresponding points, while preserving polygon faces and normals. The third technique used was the refinement of a polygonal approximation of a smooth surface. Here the number of polygons is increased as the data value increases. Initial results show, that for a small number (three to four) of distinct shapes, the turbulated cube and the refinement of a circle allow very good discrimination of shape. A detailed study of the information space would allow up to ten distinct shapes. However, this approach to representing data is limited by the perceptual limitations of the human visual system.

The proposal for the paper.

A link to some of the images .

pradyut panda
Last modified: Thu Jul 16 12:04:13 EDT 1998